From 40ad275992fec7d43a2507c768d4efdf8cbfa43c Mon Sep 17 00:00:00 2001 From: mhjensen Date: Wed, 27 Nov 2019 22:19:39 +0100 Subject: [PATCH] added material here --- doc/pub/summary/html/._summary-bs000.html | 73 ++-- doc/pub/summary/html/._summary-bs001.html | 73 ++-- doc/pub/summary/html/._summary-bs002.html | 73 ++-- doc/pub/summary/html/._summary-bs003.html | 73 ++-- doc/pub/summary/html/._summary-bs004.html | 73 ++-- doc/pub/summary/html/._summary-bs005.html | 73 ++-- doc/pub/summary/html/._summary-bs006.html | 73 ++-- doc/pub/summary/html/._summary-bs007.html | 73 ++-- doc/pub/summary/html/._summary-bs008.html | 73 ++-- doc/pub/summary/html/._summary-bs009.html | 85 +++-- doc/pub/summary/html/summary-bs.html | 73 ++-- doc/pub/summary/html/summary-reveal.html | 317 ++++++++++++++-- doc/pub/summary/html/summary-solarized.html | 353 ++++++++++++++++-- doc/pub/summary/html/summary.html | 353 ++++++++++++++++-- .../summary/ipynb/ipynb-summary-src.tar.gz | Bin 2872413 -> 2872413 bytes doc/pub/summary/ipynb/summary.ipynb | 247 +++++++++++- doc/pub/summary/pdf/summary-minted.pdf | Bin 3125045 -> 3138698 bytes doc/src/Summary/summary.do.txt | 251 ++++++++++++- 18 files changed, 1973 insertions(+), 363 deletions(-) diff --git a/doc/pub/summary/html/._summary-bs000.html b/doc/pub/summary/html/._summary-bs000.html index 405e7e297..b9bf38dd4 100644 --- a/doc/pub/summary/html/._summary-bs000.html +++ b/doc/pub/summary/html/._summary-bs000.html @@ -58,26 +58,43 @@ Automatically generated HTML file from DocOnce source '___sec5'), ('Perspective on Machine Learning', 2, None, '___sec6'), ('Machine Learning Research', 2, None, '___sec7'), - ('Hot Topics Now', 2, None, '___sec8'), - ('Starting your Machine Learning Project', 2, None, '___sec9'), - ('Choose a Model and Algorithm', 2, None, '___sec10'), - ('Preparing Your Data', 2, None, '___sec11'), + ('Starting your Machine Learning Project', 2, None, '___sec8'), + ('Choose a Model and Algorithm', 2, None, '___sec9'), + ('Preparing Your Data', 2, None, '___sec10'), ('Which Activation and Weights to Choose in Neural Networks', 2, None, - '___sec12'), - ('Optimization Methods and Hyperparameters', 2, None, '___sec13'), - ('Resampling', 2, None, '___sec14'), + '___sec11'), + ('Optimization Methods and Hyperparameters', 2, None, '___sec12'), + ('Resampling', 2, None, '___sec13'), ('Other courses on Data science and Machine Learning at UiO', 2, None, - '___sec15'), - ('Additional courses of interest', 2, None, '___sec16'), + '___sec14'), + ('Additional courses of interest', 2, None, '___sec15'), + ('Hot Topics Now', 2, None, '___sec16'), + ('Reinforcement Learning', 2, None, '___sec17'), + ('Transfer learning', 2, None, '___sec18'), + ('Adversarial learning', 2, None, '___sec19'), + ('Dual learning', 2, None, '___sec20'), + ('Distributed machine learning', 2, None, '___sec21'), + ('Meta learning', 2, None, '___sec22'), + ('The Challenges Facing Machine Learning', 2, None, '___sec23'), + ('Explainable machine learning', 2, None, '___sec24'), + ('Quantum machine learning', 2, None, '___sec25'), + ('Quantum machine learning algorithms based on linear algebra', + 2, + None, + '___sec26'), + ('Quantum reinforcement learning', 2, None, '___sec27'), + ('Quantum deep learning', 2, None, '___sec28'), + ('Social machine learning', 2, None, '___sec29'), + ('The last words?', 2, None, '___sec30'), ('Best wishes to you all and thanks so much for your heroic ' 'efforts this semester', 2, None, - '___sec17')]} + '___sec31')]} end of tocinfo --> @@ -123,16 +140,30 @@ MathJax.Hub.Config({
  • Learning outcomes and overarching aims of this course
  • Perspective on Machine Learning
  • Machine Learning Research
    • -
  • Hot Topics Now
    • -
  • Starting your Machine Learning Project
    • -
  • Choose a Model and Algorithm
    • -
  • Preparing Your Data
    • -
  • Which Activation and Weights to Choose in Neural Networks
    • -
  • Optimization Methods and Hyperparameters
    • -
  • Resampling
    • -
  • Other courses on Data science and Machine Learning at UiO
    • -
  • Additional courses of interest
    • -
  • Best wishes to you all and thanks so much for your heroic efforts this semester
    • +
  • Starting your Machine Learning Project
    • +
  • Choose a Model and Algorithm
    • +
  • Preparing Your Data
    • +
  • Which Activation and Weights to Choose in Neural Networks
    • +
  • Optimization Methods and Hyperparameters
    • +
  • Resampling
    • +
  • Other courses on Data science and Machine Learning at UiO
    • +
  • Additional courses of interest
    • +
  • Hot Topics Now
    • +
  • Reinforcement Learning
    • +
  • Transfer learning
    • +
  • Adversarial learning
    • +
  • Dual learning
    • +
  • Distributed machine learning
    • +
  • Meta learning
    • +
  • The Challenges Facing Machine Learning
    • +
  • Explainable machine learning
    • +
  • Quantum machine learning
    • +
  • Quantum machine learning algorithms based on linear algebra
    • +
  • Quantum reinforcement learning
    • +
  • Quantum deep learning
    • +
  • Social machine learning
    • +
  • The last words?
    • +
  • Best wishes to you all and thanks so much for your heroic efforts this semester
    • @@ -191,7 +222,7 @@ MathJax.Hub.Config({
  • 9
  • 10
  • ...
    • -
  • 19
    • +
  • 33
  • »
    • diff --git a/doc/pub/summary/html/._summary-bs001.html b/doc/pub/summary/html/._summary-bs001.html index 152fba497..d8d6ebbf7 100644 --- a/doc/pub/summary/html/._summary-bs001.html +++ b/doc/pub/summary/html/._summary-bs001.html @@ -58,26 +58,43 @@ Automatically generated HTML file from DocOnce source '___sec5'), ('Perspective on Machine Learning', 2, None, '___sec6'), ('Machine Learning Research', 2, None, '___sec7'), - ('Hot Topics Now', 2, None, '___sec8'), - ('Starting your Machine Learning Project', 2, None, '___sec9'), - ('Choose a Model and Algorithm', 2, None, '___sec10'), - ('Preparing Your Data', 2, None, '___sec11'), + ('Starting your Machine Learning Project', 2, None, '___sec8'), + ('Choose a Model and Algorithm', 2, None, '___sec9'), + ('Preparing Your Data', 2, None, '___sec10'), ('Which Activation and Weights to Choose in Neural Networks', 2, None, - '___sec12'), - ('Optimization Methods and Hyperparameters', 2, None, '___sec13'), - ('Resampling', 2, None, '___sec14'), + '___sec11'), + ('Optimization Methods and Hyperparameters', 2, None, '___sec12'), + ('Resampling', 2, None, '___sec13'), ('Other courses on Data science and Machine Learning at UiO', 2, None, - '___sec15'), - ('Additional courses of interest', 2, None, '___sec16'), + '___sec14'), + ('Additional courses of interest', 2, None, '___sec15'), + ('Hot Topics Now', 2, None, '___sec16'), + ('Reinforcement Learning', 2, None, '___sec17'), + ('Transfer learning', 2, None, '___sec18'), + ('Adversarial learning', 2, None, '___sec19'), + ('Dual learning', 2, None, '___sec20'), + ('Distributed machine learning', 2, None, '___sec21'), + ('Meta learning', 2, None, '___sec22'), + ('The Challenges Facing Machine Learning', 2, None, '___sec23'), + ('Explainable machine learning', 2, None, '___sec24'), + ('Quantum machine learning', 2, None, '___sec25'), + ('Quantum machine learning algorithms based on linear algebra', + 2, + None, + '___sec26'), + ('Quantum reinforcement learning', 2, None, '___sec27'), + ('Quantum deep learning', 2, None, '___sec28'), + ('Social machine learning', 2, None, '___sec29'), + ('The last words?', 2, None, '___sec30'), ('Best wishes to you all and thanks so much for your heroic ' 'efforts this semester', 2, None, - '___sec17')]} + '___sec31')]} end of tocinfo --> @@ -123,16 +140,30 @@ MathJax.Hub.Config({
  • Learning outcomes and overarching aims of this course
  • Perspective on Machine Learning
  • Machine Learning Research
    • -
  • Hot Topics Now
    • -
  • Starting your Machine Learning Project
    • -
  • Choose a Model and Algorithm
    • -
  • Preparing Your Data
    • -
  • Which Activation and Weights to Choose in Neural Networks
    • -
  • Optimization Methods and Hyperparameters
    • -
  • Resampling
    • -
  • Other courses on Data science and Machine Learning at UiO
    • -
  • Additional courses of interest
    • -
  • Best wishes to you all and thanks so much for your heroic efforts this semester
    • +
  • Starting your Machine Learning Project
    • +
  • Choose a Model and Algorithm
    • +
  • Preparing Your Data
    • +
  • Which Activation and Weights to Choose in Neural Networks
    • +
  • Optimization Methods and Hyperparameters
    • +
  • Resampling
    • +
  • Other courses on Data science and Machine Learning at UiO
    • +
  • Additional courses of interest
    • +
  • Hot Topics Now
    • +
  • Reinforcement Learning
    • +
  • Transfer learning
    • +
  • Adversarial learning
    • +
  • Dual learning
    • +
  • Distributed machine learning
    • +
  • Meta learning
    • +
  • The Challenges Facing Machine Learning
    • +
  • Explainable machine learning
    • +
  • Quantum machine learning
    • +
  • Quantum machine learning algorithms based on linear algebra
    • +
  • Quantum reinforcement learning
    • +
  • Quantum deep learning
    • +
  • Social machine learning
    • +
  • The last words?
    • +
  • Best wishes to you all and thanks so much for your heroic efforts this semester
    • @@ -169,7 +200,7 @@ MathJax.Hub.Config({
  • 10
  • 11
  • ...
    • -
  • 19
    • +
  • 33
  • »
    • diff --git a/doc/pub/summary/html/._summary-bs002.html b/doc/pub/summary/html/._summary-bs002.html index 1052817e8..7fae2379c 100644 --- a/doc/pub/summary/html/._summary-bs002.html +++ b/doc/pub/summary/html/._summary-bs002.html @@ -58,26 +58,43 @@ Automatically generated HTML file from DocOnce source '___sec5'), ('Perspective on Machine Learning', 2, None, '___sec6'), ('Machine Learning Research', 2, None, '___sec7'), - ('Hot Topics Now', 2, None, '___sec8'), - ('Starting your Machine Learning Project', 2, None, '___sec9'), - ('Choose a Model and Algorithm', 2, None, '___sec10'), - ('Preparing Your Data', 2, None, '___sec11'), + ('Starting your Machine Learning Project', 2, None, '___sec8'), + ('Choose a Model and Algorithm', 2, None, '___sec9'), + ('Preparing Your Data', 2, None, '___sec10'), ('Which Activation and Weights to Choose in Neural Networks', 2, None, - '___sec12'), - ('Optimization Methods and Hyperparameters', 2, None, '___sec13'), - ('Resampling', 2, None, '___sec14'), + '___sec11'), + ('Optimization Methods and Hyperparameters', 2, None, '___sec12'), + ('Resampling', 2, None, '___sec13'), ('Other courses on Data science and Machine Learning at UiO', 2, None, - '___sec15'), - ('Additional courses of interest', 2, None, '___sec16'), + '___sec14'), + ('Additional courses of interest', 2, None, '___sec15'), + ('Hot Topics Now', 2, None, '___sec16'), + ('Reinforcement Learning', 2, None, '___sec17'), + ('Transfer learning', 2, None, '___sec18'), + ('Adversarial learning', 2, None, '___sec19'), + ('Dual learning', 2, None, '___sec20'), + ('Distributed machine learning', 2, None, '___sec21'), + ('Meta learning', 2, None, '___sec22'), + ('The Challenges Facing Machine Learning', 2, None, '___sec23'), + ('Explainable machine learning', 2, None, '___sec24'), + ('Quantum machine learning', 2, None, '___sec25'), + ('Quantum machine learning algorithms based on linear algebra', + 2, + None, + '___sec26'), + ('Quantum reinforcement learning', 2, None, '___sec27'), + ('Quantum deep learning', 2, None, '___sec28'), + ('Social machine learning', 2, None, '___sec29'), + ('The last words?', 2, None, '___sec30'), ('Best wishes to you all and thanks so much for your heroic ' 'efforts this semester', 2, None, - '___sec17')]} + '___sec31')]} end of tocinfo --> @@ -123,16 +140,30 @@ MathJax.Hub.Config({
  • Learning outcomes and overarching aims of this course
  • Perspective on Machine Learning
  • Machine Learning Research
    • -
  • Hot Topics Now
    • -
  • Starting your Machine Learning Project
    • -
  • Choose a Model and Algorithm
    • -
  • Preparing Your Data
    • -
  • Which Activation and Weights to Choose in Neural Networks
    • -
  • Optimization Methods and Hyperparameters
    • -
  • Resampling
    • -
  • Other courses on Data science and Machine Learning at UiO
    • -
  • Additional courses of interest
    • -
  • Best wishes to you all and thanks so much for your heroic efforts this semester
    • +
  • Starting your Machine Learning Project
    • +
  • Choose a Model and Algorithm
    • +
  • Preparing Your Data
    • +
  • Which Activation and Weights to Choose in Neural Networks
    • +
  • Optimization Methods and Hyperparameters
    • +
  • Resampling
    • +
  • Other courses on Data science and Machine Learning at UiO
    • +
  • Additional courses of interest
    • +
  • Hot Topics Now
    • +
  • Reinforcement Learning
    • +
  • Transfer learning
    • +
  • Adversarial learning
    • +
  • Dual learning
    • +
  • Distributed machine learning
    • +
  • Meta learning
    • +
  • The Challenges Facing Machine Learning
    • +
  • Explainable machine learning
    • +
  • Quantum machine learning
    • +
  • Quantum machine learning algorithms based on linear algebra
    • +
  • Quantum reinforcement learning
    • +
  • Quantum deep learning
    • +
  • Social machine learning
    • +
  • The last words?
    • +
  • Best wishes to you all and thanks so much for your heroic efforts this semester
    • @@ -175,7 +206,7 @@ Does that match the experiences you have made this semester?
  • 11
  • 12
  • ...
    • -
  • 19
    • +
  • 33
  • »
    • diff --git a/doc/pub/summary/html/._summary-bs003.html b/doc/pub/summary/html/._summary-bs003.html index c41f7a305..d38ca7df0 100644 --- a/doc/pub/summary/html/._summary-bs003.html +++ b/doc/pub/summary/html/._summary-bs003.html @@ -58,26 +58,43 @@ Automatically generated HTML file from DocOnce source '___sec5'), ('Perspective on Machine Learning', 2, None, '___sec6'), ('Machine Learning Research', 2, None, '___sec7'), - ('Hot Topics Now', 2, None, '___sec8'), - ('Starting your Machine Learning Project', 2, None, '___sec9'), - ('Choose a Model and Algorithm', 2, None, '___sec10'), - ('Preparing Your Data', 2, None, '___sec11'), + ('Starting your Machine Learning Project', 2, None, '___sec8'), + ('Choose a Model and Algorithm', 2, None, '___sec9'), + ('Preparing Your Data', 2, None, '___sec10'), ('Which Activation and Weights to Choose in Neural Networks', 2, None, - '___sec12'), - ('Optimization Methods and Hyperparameters', 2, None, '___sec13'), - ('Resampling', 2, None, '___sec14'), + '___sec11'), + ('Optimization Methods and Hyperparameters', 2, None, '___sec12'), + ('Resampling', 2, None, '___sec13'), ('Other courses on Data science and Machine Learning at UiO', 2, None, - '___sec15'), - ('Additional courses of interest', 2, None, '___sec16'), + '___sec14'), + ('Additional courses of interest', 2, None, '___sec15'), + ('Hot Topics Now', 2, None, '___sec16'), + ('Reinforcement Learning', 2, None, '___sec17'), + ('Transfer learning', 2, None, '___sec18'), + ('Adversarial learning', 2, None, '___sec19'), + ('Dual learning', 2, None, '___sec20'), + ('Distributed machine learning', 2, None, '___sec21'), + ('Meta learning', 2, None, '___sec22'), + ('The Challenges Facing Machine Learning', 2, None, '___sec23'), + ('Explainable machine learning', 2, None, '___sec24'), + ('Quantum machine learning', 2, None, '___sec25'), + ('Quantum machine learning algorithms based on linear algebra', + 2, + None, + '___sec26'), + ('Quantum reinforcement learning', 2, None, '___sec27'), + ('Quantum deep learning', 2, None, '___sec28'), + ('Social machine learning', 2, None, '___sec29'), + ('The last words?', 2, None, '___sec30'), ('Best wishes to you all and thanks so much for your heroic ' 'efforts this semester', 2, None, - '___sec17')]} + '___sec31')]} end of tocinfo --> @@ -123,16 +140,30 @@ MathJax.Hub.Config({
  • Learning outcomes and overarching aims of this course
  • Perspective on Machine Learning
  • Machine Learning Research
    • -
  • Hot Topics Now
    • -
  • Starting your Machine Learning Project
    • -
  • Choose a Model and Algorithm
    • -
  • Preparing Your Data
    • -
  • Which Activation and Weights to Choose in Neural Networks
    • -
  • Optimization Methods and Hyperparameters
    • -
  • Resampling
    • -
  • Other courses on Data science and Machine Learning at UiO
    • -
  • Additional courses of interest
    • -
  • Best wishes to you all and thanks so much for your heroic efforts this semester
    • +
  • Starting your Machine Learning Project
    • +
  • Choose a Model and Algorithm
    • +
  • Preparing Your Data
    • +
  • Which Activation and Weights to Choose in Neural Networks
    • +
  • Optimization Methods and Hyperparameters
    • +
  • Resampling
    • +
  • Other courses on Data science and Machine Learning at UiO
    • +
  • Additional courses of interest
    • +
  • Hot Topics Now
    • +
  • Reinforcement Learning
    • +
  • Transfer learning
    • +
  • Adversarial learning
    • +
  • Dual learning
    • +
  • Distributed machine learning
    • +
  • Meta learning
    • +
  • The Challenges Facing Machine Learning
    • +
  • Explainable machine learning
    • +
  • Quantum machine learning
    • +
  • Quantum machine learning algorithms based on linear algebra
    • +
  • Quantum reinforcement learning
    • +
  • Quantum deep learning
    • +
  • Social machine learning
    • +
  • The last words?
    • +
  • Best wishes to you all and thanks so much for your heroic efforts this semester
    • @@ -176,7 +207,7 @@ The course has two central parts
  • 12
  • 13
  • ...
    • -
  • 19
    • +
  • 33
  • »
    • diff --git a/doc/pub/summary/html/._summary-bs004.html b/doc/pub/summary/html/._summary-bs004.html index a0124c23c..c836ba82e 100644 --- a/doc/pub/summary/html/._summary-bs004.html +++ b/doc/pub/summary/html/._summary-bs004.html @@ -58,26 +58,43 @@ Automatically generated HTML file from DocOnce source '___sec5'), ('Perspective on Machine Learning', 2, None, '___sec6'), ('Machine Learning Research', 2, None, '___sec7'), - ('Hot Topics Now', 2, None, '___sec8'), - ('Starting your Machine Learning Project', 2, None, '___sec9'), - ('Choose a Model and Algorithm', 2, None, '___sec10'), - ('Preparing Your Data', 2, None, '___sec11'), + ('Starting your Machine Learning Project', 2, None, '___sec8'), + ('Choose a Model and Algorithm', 2, None, '___sec9'), + ('Preparing Your Data', 2, None, '___sec10'), ('Which Activation and Weights to Choose in Neural Networks', 2, None, - '___sec12'), - ('Optimization Methods and Hyperparameters', 2, None, '___sec13'), - ('Resampling', 2, None, '___sec14'), + '___sec11'), + ('Optimization Methods and Hyperparameters', 2, None, '___sec12'), + ('Resampling', 2, None, '___sec13'), ('Other courses on Data science and Machine Learning at UiO', 2, None, - '___sec15'), - ('Additional courses of interest', 2, None, '___sec16'), + '___sec14'), + ('Additional courses of interest', 2, None, '___sec15'), + ('Hot Topics Now', 2, None, '___sec16'), + ('Reinforcement Learning', 2, None, '___sec17'), + ('Transfer learning', 2, None, '___sec18'), + ('Adversarial learning', 2, None, '___sec19'), + ('Dual learning', 2, None, '___sec20'), + ('Distributed machine learning', 2, None, '___sec21'), + ('Meta learning', 2, None, '___sec22'), + ('The Challenges Facing Machine Learning', 2, None, '___sec23'), + ('Explainable machine learning', 2, None, '___sec24'), + ('Quantum machine learning', 2, None, '___sec25'), + ('Quantum machine learning algorithms based on linear algebra', + 2, + None, + '___sec26'), + ('Quantum reinforcement learning', 2, None, '___sec27'), + ('Quantum deep learning', 2, None, '___sec28'), + ('Social machine learning', 2, None, '___sec29'), + ('The last words?', 2, None, '___sec30'), ('Best wishes to you all and thanks so much for your heroic ' 'efforts this semester', 2, None, - '___sec17')]} + '___sec31')]} end of tocinfo --> @@ -123,16 +140,30 @@ MathJax.Hub.Config({
  • Learning outcomes and overarching aims of this course
  • Perspective on Machine Learning
  • Machine Learning Research
    • -
  • Hot Topics Now
    • -
  • Starting your Machine Learning Project
    • -
  • Choose a Model and Algorithm
    • -
  • Preparing Your Data
    • -
  • Which Activation and Weights to Choose in Neural Networks
    • -
  • Optimization Methods and Hyperparameters
    • -
  • Resampling
    • -
  • Other courses on Data science and Machine Learning at UiO
    • -
  • Additional courses of interest
    • -
  • Best wishes to you all and thanks so much for your heroic efforts this semester
    • +
  • Starting your Machine Learning Project
    • +
  • Choose a Model and Algorithm
    • +
  • Preparing Your Data
    • +
  • Which Activation and Weights to Choose in Neural Networks
    • +
  • Optimization Methods and Hyperparameters
    • +
  • Resampling
    • +
  • Other courses on Data science and Machine Learning at UiO
    • +
  • Additional courses of interest
    • +
  • Hot Topics Now
    • +
  • Reinforcement Learning
    • +
  • Transfer learning
    • +
  • Adversarial learning
    • +
  • Dual learning
    • +
  • Distributed machine learning
    • +
  • Meta learning
    • +
  • The Challenges Facing Machine Learning
    • +
  • Explainable machine learning
    • +
  • Quantum machine learning
    • +
  • Quantum machine learning algorithms based on linear algebra
    • +
  • Quantum reinforcement learning
    • +
  • Quantum deep learning
    • +
  • Social machine learning
    • +
  • The last words?
    • +
  • Best wishes to you all and thanks so much for your heroic efforts this semester
    • @@ -183,7 +214,7 @@ The following topics will be covered
  • 13
  • 14
  • ...
    • -
  • 19
    • +
  • 33
  • »
    • diff --git a/doc/pub/summary/html/._summary-bs005.html b/doc/pub/summary/html/._summary-bs005.html index 0ebfa4c59..c78331d17 100644 --- a/doc/pub/summary/html/._summary-bs005.html +++ b/doc/pub/summary/html/._summary-bs005.html @@ -58,26 +58,43 @@ Automatically generated HTML file from DocOnce source '___sec5'), ('Perspective on Machine Learning', 2, None, '___sec6'), ('Machine Learning Research', 2, None, '___sec7'), - ('Hot Topics Now', 2, None, '___sec8'), - ('Starting your Machine Learning Project', 2, None, '___sec9'), - ('Choose a Model and Algorithm', 2, None, '___sec10'), - ('Preparing Your Data', 2, None, '___sec11'), + ('Starting your Machine Learning Project', 2, None, '___sec8'), + ('Choose a Model and Algorithm', 2, None, '___sec9'), + ('Preparing Your Data', 2, None, '___sec10'), ('Which Activation and Weights to Choose in Neural Networks', 2, None, - '___sec12'), - ('Optimization Methods and Hyperparameters', 2, None, '___sec13'), - ('Resampling', 2, None, '___sec14'), + '___sec11'), + ('Optimization Methods and Hyperparameters', 2, None, '___sec12'), + ('Resampling', 2, None, '___sec13'), ('Other courses on Data science and Machine Learning at UiO', 2, None, - '___sec15'), - ('Additional courses of interest', 2, None, '___sec16'), + '___sec14'), + ('Additional courses of interest', 2, None, '___sec15'), + ('Hot Topics Now', 2, None, '___sec16'), + ('Reinforcement Learning', 2, None, '___sec17'), + ('Transfer learning', 2, None, '___sec18'), + ('Adversarial learning', 2, None, '___sec19'), + ('Dual learning', 2, None, '___sec20'), + ('Distributed machine learning', 2, None, '___sec21'), + ('Meta learning', 2, None, '___sec22'), + ('The Challenges Facing Machine Learning', 2, None, '___sec23'), + ('Explainable machine learning', 2, None, '___sec24'), + ('Quantum machine learning', 2, None, '___sec25'), + ('Quantum machine learning algorithms based on linear algebra', + 2, + None, + '___sec26'), + ('Quantum reinforcement learning', 2, None, '___sec27'), + ('Quantum deep learning', 2, None, '___sec28'), + ('Social machine learning', 2, None, '___sec29'), + ('The last words?', 2, None, '___sec30'), ('Best wishes to you all and thanks so much for your heroic ' 'efforts this semester', 2, None, - '___sec17')]} + '___sec31')]} end of tocinfo --> @@ -123,16 +140,30 @@ MathJax.Hub.Config({
  • Learning outcomes and overarching aims of this course
  • Perspective on Machine Learning
  • Machine Learning Research
    • -
  • Hot Topics Now
    • -
  • Starting your Machine Learning Project
    • -
  • Choose a Model and Algorithm
    • -
  • Preparing Your Data
    • -
  • Which Activation and Weights to Choose in Neural Networks
    • -
  • Optimization Methods and Hyperparameters
    • -
  • Resampling
    • -
  • Other courses on Data science and Machine Learning at UiO
    • -
  • Additional courses of interest
    • -
  • Best wishes to you all and thanks so much for your heroic efforts this semester
    • +
  • Starting your Machine Learning Project
    • +
  • Choose a Model and Algorithm
    • +
  • Preparing Your Data
    • +
  • Which Activation and Weights to Choose in Neural Networks
    • +
  • Optimization Methods and Hyperparameters
    • +
  • Resampling
    • +
  • Other courses on Data science and Machine Learning at UiO
    • +
  • Additional courses of interest
    • +
  • Hot Topics Now
    • +
  • Reinforcement Learning
    • +
  • Transfer learning
    • +
  • Adversarial learning
    • +
  • Dual learning
    • +
  • Distributed machine learning
    • +
  • Meta learning
    • +
  • The Challenges Facing Machine Learning
    • +
  • Explainable machine learning
    • +
  • Quantum machine learning
    • +
  • Quantum machine learning algorithms based on linear algebra
    • +
  • Quantum reinforcement learning
    • +
  • Quantum deep learning
    • +
  • Social machine learning
    • +
  • The last words?
    • +
  • Best wishes to you all and thanks so much for your heroic efforts this semester
    • @@ -180,7 +211,7 @@ The following topics will be covered
  • 14
  • 15
  • ...
    • -
  • 19
    • +
  • 33
  • »
    • diff --git a/doc/pub/summary/html/._summary-bs006.html b/doc/pub/summary/html/._summary-bs006.html index 86b833c1e..82f64abb5 100644 --- a/doc/pub/summary/html/._summary-bs006.html +++ b/doc/pub/summary/html/._summary-bs006.html @@ -58,26 +58,43 @@ Automatically generated HTML file from DocOnce source '___sec5'), ('Perspective on Machine Learning', 2, None, '___sec6'), ('Machine Learning Research', 2, None, '___sec7'), - ('Hot Topics Now', 2, None, '___sec8'), - ('Starting your Machine Learning Project', 2, None, '___sec9'), - ('Choose a Model and Algorithm', 2, None, '___sec10'), - ('Preparing Your Data', 2, None, '___sec11'), + ('Starting your Machine Learning Project', 2, None, '___sec8'), + ('Choose a Model and Algorithm', 2, None, '___sec9'), + ('Preparing Your Data', 2, None, '___sec10'), ('Which Activation and Weights to Choose in Neural Networks', 2, None, - '___sec12'), - ('Optimization Methods and Hyperparameters', 2, None, '___sec13'), - ('Resampling', 2, None, '___sec14'), + '___sec11'), + ('Optimization Methods and Hyperparameters', 2, None, '___sec12'), + ('Resampling', 2, None, '___sec13'), ('Other courses on Data science and Machine Learning at UiO', 2, None, - '___sec15'), - ('Additional courses of interest', 2, None, '___sec16'), + '___sec14'), + ('Additional courses of interest', 2, None, '___sec15'), + ('Hot Topics Now', 2, None, '___sec16'), + ('Reinforcement Learning', 2, None, '___sec17'), + ('Transfer learning', 2, None, '___sec18'), + ('Adversarial learning', 2, None, '___sec19'), + ('Dual learning', 2, None, '___sec20'), + ('Distributed machine learning', 2, None, '___sec21'), + ('Meta learning', 2, None, '___sec22'), + ('The Challenges Facing Machine Learning', 2, None, '___sec23'), + ('Explainable machine learning', 2, None, '___sec24'), + ('Quantum machine learning', 2, None, '___sec25'), + ('Quantum machine learning algorithms based on linear algebra', + 2, + None, + '___sec26'), + ('Quantum reinforcement learning', 2, None, '___sec27'), + ('Quantum deep learning', 2, None, '___sec28'), + ('Social machine learning', 2, None, '___sec29'), + ('The last words?', 2, None, '___sec30'), ('Best wishes to you all and thanks so much for your heroic ' 'efforts this semester', 2, None, - '___sec17')]} + '___sec31')]} end of tocinfo --> @@ -123,16 +140,30 @@ MathJax.Hub.Config({
  • Learning outcomes and overarching aims of this course
  • Perspective on Machine Learning
  • Machine Learning Research
    • -
  • Hot Topics Now
    • -
  • Starting your Machine Learning Project
    • -
  • Choose a Model and Algorithm
    • -
  • Preparing Your Data
    • -
  • Which Activation and Weights to Choose in Neural Networks
    • -
  • Optimization Methods and Hyperparameters
    • -
  • Resampling
    • -
  • Other courses on Data science and Machine Learning at UiO
    • -
  • Additional courses of interest
    • -
  • Best wishes to you all and thanks so much for your heroic efforts this semester
    • +
  • Starting your Machine Learning Project
    • +
  • Choose a Model and Algorithm
    • +
  • Preparing Your Data
    • +
  • Which Activation and Weights to Choose in Neural Networks
    • +
  • Optimization Methods and Hyperparameters
    • +
  • Resampling
    • +
  • Other courses on Data science and Machine Learning at UiO
    • +
  • Additional courses of interest
    • +
  • Hot Topics Now
    • +
  • Reinforcement Learning
    • +
  • Transfer learning
    • +
  • Adversarial learning
    • +
  • Dual learning
    • +
  • Distributed machine learning
    • +
  • Meta learning
    • +
  • The Challenges Facing Machine Learning
    • +
  • Explainable machine learning
    • +
  • Quantum machine learning
    • +
  • Quantum machine learning algorithms based on linear algebra
    • +
  • Quantum reinforcement learning
    • +
  • Quantum deep learning
    • +
  • Social machine learning
    • +
  • The last words?
    • +
  • Best wishes to you all and thanks so much for your heroic efforts this semester
    • @@ -193,7 +224,7 @@ ethical conduct is emphasized throughout the course.
  • 15
  • 16
  • ...
    • -
  • 19
    • +
  • 33
  • »
    • diff --git a/doc/pub/summary/html/._summary-bs007.html b/doc/pub/summary/html/._summary-bs007.html index b9242504a..d83a0f69a 100644 --- a/doc/pub/summary/html/._summary-bs007.html +++ b/doc/pub/summary/html/._summary-bs007.html @@ -58,26 +58,43 @@ Automatically generated HTML file from DocOnce source '___sec5'), ('Perspective on Machine Learning', 2, None, '___sec6'), ('Machine Learning Research', 2, None, '___sec7'), - ('Hot Topics Now', 2, None, '___sec8'), - ('Starting your Machine Learning Project', 2, None, '___sec9'), - ('Choose a Model and Algorithm', 2, None, '___sec10'), - ('Preparing Your Data', 2, None, '___sec11'), + ('Starting your Machine Learning Project', 2, None, '___sec8'), + ('Choose a Model and Algorithm', 2, None, '___sec9'), + ('Preparing Your Data', 2, None, '___sec10'), ('Which Activation and Weights to Choose in Neural Networks', 2, None, - '___sec12'), - ('Optimization Methods and Hyperparameters', 2, None, '___sec13'), - ('Resampling', 2, None, '___sec14'), + '___sec11'), + ('Optimization Methods and Hyperparameters', 2, None, '___sec12'), + ('Resampling', 2, None, '___sec13'), ('Other courses on Data science and Machine Learning at UiO', 2, None, - '___sec15'), - ('Additional courses of interest', 2, None, '___sec16'), + '___sec14'), + ('Additional courses of interest', 2, None, '___sec15'), + ('Hot Topics Now', 2, None, '___sec16'), + ('Reinforcement Learning', 2, None, '___sec17'), + ('Transfer learning', 2, None, '___sec18'), + ('Adversarial learning', 2, None, '___sec19'), + ('Dual learning', 2, None, '___sec20'), + ('Distributed machine learning', 2, None, '___sec21'), + ('Meta learning', 2, None, '___sec22'), + ('The Challenges Facing Machine Learning', 2, None, '___sec23'), + ('Explainable machine learning', 2, None, '___sec24'), + ('Quantum machine learning', 2, None, '___sec25'), + ('Quantum machine learning algorithms based on linear algebra', + 2, + None, + '___sec26'), + ('Quantum reinforcement learning', 2, None, '___sec27'), + ('Quantum deep learning', 2, None, '___sec28'), + ('Social machine learning', 2, None, '___sec29'), + ('The last words?', 2, None, '___sec30'), ('Best wishes to you all and thanks so much for your heroic ' 'efforts this semester', 2, None, - '___sec17')]} + '___sec31')]} end of tocinfo --> @@ -123,16 +140,30 @@ MathJax.Hub.Config({
  • Learning outcomes and overarching aims of this course
  • Perspective on Machine Learning
  • Machine Learning Research
    • -
  • Hot Topics Now
    • -
  • Starting your Machine Learning Project
    • -
  • Choose a Model and Algorithm
    • -
  • Preparing Your Data
    • -
  • Which Activation and Weights to Choose in Neural Networks
    • -
  • Optimization Methods and Hyperparameters
    • -
  • Resampling
    • -
  • Other courses on Data science and Machine Learning at UiO
    • -
  • Additional courses of interest
    • -
  • Best wishes to you all and thanks so much for your heroic efforts this semester
    • +
  • Starting your Machine Learning Project
    • +
  • Choose a Model and Algorithm
    • +
  • Preparing Your Data
    • +
  • Which Activation and Weights to Choose in Neural Networks
    • +
  • Optimization Methods and Hyperparameters
    • +
  • Resampling
    • +
  • Other courses on Data science and Machine Learning at UiO
    • +
  • Additional courses of interest
    • +
  • Hot Topics Now
    • +
  • Reinforcement Learning
    • +
  • Transfer learning
    • +
  • Adversarial learning
    • +
  • Dual learning
    • +
  • Distributed machine learning
    • +
  • Meta learning
    • +
  • The Challenges Facing Machine Learning
    • +
  • Explainable machine learning
    • +
  • Quantum machine learning
    • +
  • Quantum machine learning algorithms based on linear algebra
    • +
  • Quantum reinforcement learning
    • +
  • Quantum deep learning
    • +
  • Social machine learning
    • +
  • The last words?
    • +
  • Best wishes to you all and thanks so much for your heroic efforts this semester
    • @@ -185,7 +216,7 @@ Neural Networks, etc.
  • 16
  • 17
  • ...
    • -
  • 19
    • +
  • 33
  • »
    • diff --git a/doc/pub/summary/html/._summary-bs008.html b/doc/pub/summary/html/._summary-bs008.html index 2cd673b49..69b5b39d7 100644 --- a/doc/pub/summary/html/._summary-bs008.html +++ b/doc/pub/summary/html/._summary-bs008.html @@ -58,26 +58,43 @@ Automatically generated HTML file from DocOnce source '___sec5'), ('Perspective on Machine Learning', 2, None, '___sec6'), ('Machine Learning Research', 2, None, '___sec7'), - ('Hot Topics Now', 2, None, '___sec8'), - ('Starting your Machine Learning Project', 2, None, '___sec9'), - ('Choose a Model and Algorithm', 2, None, '___sec10'), - ('Preparing Your Data', 2, None, '___sec11'), + ('Starting your Machine Learning Project', 2, None, '___sec8'), + ('Choose a Model and Algorithm', 2, None, '___sec9'), + ('Preparing Your Data', 2, None, '___sec10'), ('Which Activation and Weights to Choose in Neural Networks', 2, None, - '___sec12'), - ('Optimization Methods and Hyperparameters', 2, None, '___sec13'), - ('Resampling', 2, None, '___sec14'), + '___sec11'), + ('Optimization Methods and Hyperparameters', 2, None, '___sec12'), + ('Resampling', 2, None, '___sec13'), ('Other courses on Data science and Machine Learning at UiO', 2, None, - '___sec15'), - ('Additional courses of interest', 2, None, '___sec16'), + '___sec14'), + ('Additional courses of interest', 2, None, '___sec15'), + ('Hot Topics Now', 2, None, '___sec16'), + ('Reinforcement Learning', 2, None, '___sec17'), + ('Transfer learning', 2, None, '___sec18'), + ('Adversarial learning', 2, None, '___sec19'), + ('Dual learning', 2, None, '___sec20'), + ('Distributed machine learning', 2, None, '___sec21'), + ('Meta learning', 2, None, '___sec22'), + ('The Challenges Facing Machine Learning', 2, None, '___sec23'), + ('Explainable machine learning', 2, None, '___sec24'), + ('Quantum machine learning', 2, None, '___sec25'), + ('Quantum machine learning algorithms based on linear algebra', + 2, + None, + '___sec26'), + ('Quantum reinforcement learning', 2, None, '___sec27'), + ('Quantum deep learning', 2, None, '___sec28'), + ('Social machine learning', 2, None, '___sec29'), + ('The last words?', 2, None, '___sec30'), ('Best wishes to you all and thanks so much for your heroic ' 'efforts this semester', 2, None, - '___sec17')]} + '___sec31')]} end of tocinfo --> @@ -123,16 +140,30 @@ MathJax.Hub.Config({
  • Learning outcomes and overarching aims of this course
  • Perspective on Machine Learning
  • Machine Learning Research
    • -
  • Hot Topics Now
    • -
  • Starting your Machine Learning Project
    • -
  • Choose a Model and Algorithm
    • -
  • Preparing Your Data
    • -
  • Which Activation and Weights to Choose in Neural Networks
    • -
  • Optimization Methods and Hyperparameters
    • -
  • Resampling
    • -
  • Other courses on Data science and Machine Learning at UiO
    • -
  • Additional courses of interest
    • -
  • Best wishes to you all and thanks so much for your heroic efforts this semester
    • +
  • Starting your Machine Learning Project
    • +
  • Choose a Model and Algorithm
    • +
  • Preparing Your Data
    • +
  • Which Activation and Weights to Choose in Neural Networks
    • +
  • Optimization Methods and Hyperparameters
    • +
  • Resampling
    • +
  • Other courses on Data science and Machine Learning at UiO
    • +
  • Additional courses of interest
    • +
  • Hot Topics Now
    • +
  • Reinforcement Learning
    • +
  • Transfer learning
    • +
  • Adversarial learning
    • +
  • Dual learning
    • +
  • Distributed machine learning
    • +
  • Meta learning
    • +
  • The Challenges Facing Machine Learning
    • +
  • Explainable machine learning
    • +
  • Quantum machine learning
    • +
  • Quantum machine learning algorithms based on linear algebra
    • +
  • Quantum reinforcement learning
    • +
  • Quantum deep learning
    • +
  • Social machine learning
    • +
  • The last words?
    • +
  • Best wishes to you all and thanks so much for your heroic efforts this semester
    • @@ -184,7 +215,7 @@ Where to find recent results:
  • 17
  • 18
  • ...
    • -
  • 19
    • +
  • 33
  • »
    • diff --git a/doc/pub/summary/html/._summary-bs009.html b/doc/pub/summary/html/._summary-bs009.html index f211e4b81..eee0f9df3 100644 --- a/doc/pub/summary/html/._summary-bs009.html +++ b/doc/pub/summary/html/._summary-bs009.html @@ -58,26 +58,43 @@ Automatically generated HTML file from DocOnce source '___sec5'), ('Perspective on Machine Learning', 2, None, '___sec6'), ('Machine Learning Research', 2, None, '___sec7'), - ('Hot Topics Now', 2, None, '___sec8'), - ('Starting your Machine Learning Project', 2, None, '___sec9'), - ('Choose a Model and Algorithm', 2, None, '___sec10'), - ('Preparing Your Data', 2, None, '___sec11'), + ('Starting your Machine Learning Project', 2, None, '___sec8'), + ('Choose a Model and Algorithm', 2, None, '___sec9'), + ('Preparing Your Data', 2, None, '___sec10'), ('Which Activation and Weights to Choose in Neural Networks', 2, None, - '___sec12'), - ('Optimization Methods and Hyperparameters', 2, None, '___sec13'), - ('Resampling', 2, None, '___sec14'), + '___sec11'), + ('Optimization Methods and Hyperparameters', 2, None, '___sec12'), + ('Resampling', 2, None, '___sec13'), ('Other courses on Data science and Machine Learning at UiO', 2, None, - '___sec15'), - ('Additional courses of interest', 2, None, '___sec16'), + '___sec14'), + ('Additional courses of interest', 2, None, '___sec15'), + ('Hot Topics Now', 2, None, '___sec16'), + ('Reinforcement Learning', 2, None, '___sec17'), + ('Transfer learning', 2, None, '___sec18'), + ('Adversarial learning', 2, None, '___sec19'), + ('Dual learning', 2, None, '___sec20'), + ('Distributed machine learning', 2, None, '___sec21'), + ('Meta learning', 2, None, '___sec22'), + ('The Challenges Facing Machine Learning', 2, None, '___sec23'), + ('Explainable machine learning', 2, None, '___sec24'), + ('Quantum machine learning', 2, None, '___sec25'), + ('Quantum machine learning algorithms based on linear algebra', + 2, + None, + '___sec26'), + ('Quantum reinforcement learning', 2, None, '___sec27'), + ('Quantum deep learning', 2, None, '___sec28'), + ('Social machine learning', 2, None, '___sec29'), + ('The last words?', 2, None, '___sec30'), ('Best wishes to you all and thanks so much for your heroic ' 'efforts this semester', 2, None, - '___sec17')]} + '___sec31')]} end of tocinfo --> @@ -123,16 +140,30 @@ MathJax.Hub.Config({
  • Learning outcomes and overarching aims of this course
  • Perspective on Machine Learning
  • Machine Learning Research
    • -
  • Hot Topics Now
    • -
  • Starting your Machine Learning Project
    • -
  • Choose a Model and Algorithm
    • -
  • Preparing Your Data
    • -
  • Which Activation and Weights to Choose in Neural Networks
    • -
  • Optimization Methods and Hyperparameters
    • -
  • Resampling
    • -
  • Other courses on Data science and Machine Learning at UiO
    • -
  • Additional courses of interest
    • -
  • Best wishes to you all and thanks so much for your heroic efforts this semester
    • +
  • Starting your Machine Learning Project
    • +
  • Choose a Model and Algorithm
    • +
  • Preparing Your Data
    • +
  • Which Activation and Weights to Choose in Neural Networks
    • +
  • Optimization Methods and Hyperparameters
    • +
  • Resampling
    • +
  • Other courses on Data science and Machine Learning at UiO
    • +
  • Additional courses of interest
    • +
  • Hot Topics Now
    • +
  • Reinforcement Learning
    • +
  • Transfer learning
    • +
  • Adversarial learning
    • +
  • Dual learning
    • +
  • Distributed machine learning
    • +
  • Meta learning
    • +
  • The Challenges Facing Machine Learning
    • +
  • Explainable machine learning
    • +
  • Quantum machine learning
    • +
  • Quantum machine learning algorithms based on linear algebra
    • +
  • Quantum reinforcement learning
    • +
  • Quantum deep learning
    • +
  • Social machine learning
    • +
  • The last words?
    • +
  • Best wishes to you all and thanks so much for your heroic efforts this semester
    • @@ -148,14 +179,14 @@ MathJax.Hub.Config({ -

    Hot Topics Now

    +

    Starting your Machine Learning Project

      -
    1. Boosting techniques and complex neural networks
      2. -
    2. Adversarial examples
      3. -
    3. Zero shot learning
      4. -
    4. Transfer learning
      5. -
    5. Model interpretability
      6. +
    6. Identify problem type: classification, generation, regression
      7. +
    7. Consider your data carefully
      8. +
    8. Choose a simple model that fits 1. and 2.
      9. +
    9. Consider your data carefully again… data representation
      10. +
    10. Based on results, feedback loop to earliest possible point

    @@ -181,6 +212,8 @@ MathJax.Hub.Config({

  • 17
  • 18
  • 19
    • +
  • ...
    • +
  • 33
  • »
    • diff --git a/doc/pub/summary/html/summary-bs.html b/doc/pub/summary/html/summary-bs.html index 405e7e297..b9bf38dd4 100644 --- a/doc/pub/summary/html/summary-bs.html +++ b/doc/pub/summary/html/summary-bs.html @@ -58,26 +58,43 @@ Automatically generated HTML file from DocOnce source '___sec5'), ('Perspective on Machine Learning', 2, None, '___sec6'), ('Machine Learning Research', 2, None, '___sec7'), - ('Hot Topics Now', 2, None, '___sec8'), - ('Starting your Machine Learning Project', 2, None, '___sec9'), - ('Choose a Model and Algorithm', 2, None, '___sec10'), - ('Preparing Your Data', 2, None, '___sec11'), + ('Starting your Machine Learning Project', 2, None, '___sec8'), + ('Choose a Model and Algorithm', 2, None, '___sec9'), + ('Preparing Your Data', 2, None, '___sec10'), ('Which Activation and Weights to Choose in Neural Networks', 2, None, - '___sec12'), - ('Optimization Methods and Hyperparameters', 2, None, '___sec13'), - ('Resampling', 2, None, '___sec14'), + '___sec11'), + ('Optimization Methods and Hyperparameters', 2, None, '___sec12'), + ('Resampling', 2, None, '___sec13'), ('Other courses on Data science and Machine Learning at UiO', 2, None, - '___sec15'), - ('Additional courses of interest', 2, None, '___sec16'), + '___sec14'), + ('Additional courses of interest', 2, None, '___sec15'), + ('Hot Topics Now', 2, None, '___sec16'), + ('Reinforcement Learning', 2, None, '___sec17'), + ('Transfer learning', 2, None, '___sec18'), + ('Adversarial learning', 2, None, '___sec19'), + ('Dual learning', 2, None, '___sec20'), + ('Distributed machine learning', 2, None, '___sec21'), + ('Meta learning', 2, None, '___sec22'), + ('The Challenges Facing Machine Learning', 2, None, '___sec23'), + ('Explainable machine learning', 2, None, '___sec24'), + ('Quantum machine learning', 2, None, '___sec25'), + ('Quantum machine learning algorithms based on linear algebra', + 2, + None, + '___sec26'), + ('Quantum reinforcement learning', 2, None, '___sec27'), + ('Quantum deep learning', 2, None, '___sec28'), + ('Social machine learning', 2, None, '___sec29'), + ('The last words?', 2, None, '___sec30'), ('Best wishes to you all and thanks so much for your heroic ' 'efforts this semester', 2, None, - '___sec17')]} + '___sec31')]} end of tocinfo --> @@ -123,16 +140,30 @@ MathJax.Hub.Config({
  • Learning outcomes and overarching aims of this course
  • Perspective on Machine Learning
  • Machine Learning Research
    • -
  • Hot Topics Now
    • -
  • Starting your Machine Learning Project
    • -
  • Choose a Model and Algorithm
    • -
  • Preparing Your Data
    • -
  • Which Activation and Weights to Choose in Neural Networks
    • -
  • Optimization Methods and Hyperparameters
    • -
  • Resampling
    • -
  • Other courses on Data science and Machine Learning at UiO
    • -
  • Additional courses of interest
    • -
  • Best wishes to you all and thanks so much for your heroic efforts this semester
    • +
  • Starting your Machine Learning Project
    • +
  • Choose a Model and Algorithm
    • +
  • Preparing Your Data
    • +
  • Which Activation and Weights to Choose in Neural Networks
    • +
  • Optimization Methods and Hyperparameters
    • +
  • Resampling
    • +
  • Other courses on Data science and Machine Learning at UiO
    • +
  • Additional courses of interest
    • +
  • Hot Topics Now
    • +
  • Reinforcement Learning
    • +
  • Transfer learning
    • +
  • Adversarial learning
    • +
  • Dual learning
    • +
  • Distributed machine learning
    • +
  • Meta learning
    • +
  • The Challenges Facing Machine Learning
    • +
  • Explainable machine learning
    • +
  • Quantum machine learning
    • +
  • Quantum machine learning algorithms based on linear algebra
    • +
  • Quantum reinforcement learning
    • +
  • Quantum deep learning
    • +
  • Social machine learning
    • +
  • The last words?
    • +
  • Best wishes to you all and thanks so much for your heroic efforts this semester
    • @@ -191,7 +222,7 @@ MathJax.Hub.Config({
  • 9
  • 10
  • ...
    • -
  • 19
    • +
  • 33
  • »
    • diff --git a/doc/pub/summary/html/summary-reveal.html b/doc/pub/summary/html/summary-reveal.html index fb147576a..6eadd32f5 100644 --- a/doc/pub/summary/html/summary-reveal.html +++ b/doc/pub/summary/html/summary-reveal.html @@ -286,20 +286,7 @@ Where to find recent results:
    -

    Hot Topics Now

    - -
      -

    1. Boosting techniques and complex neural networks
      2. -

    2. Adversarial examples
      3. -

    3. Zero shot learning
      4. -

    4. Transfer learning
      5. -

    5. Model interpretability
      6. -
    -
    - - -
    -

    Starting your Machine Learning Project

    +

    Starting your Machine Learning Project

    1. Identify problem type: classification, generation, regression
      2. @@ -312,7 +299,7 @@ Where to find recent results:
      -

      Choose a Model and Algorithm

      +

      Choose a Model and Algorithm

      1. Supervised?
        2. @@ -323,7 +310,7 @@ Where to find recent results:
        -

        Preparing Your Data

        +

        Preparing Your Data

        1. Shuffle your data
          2. @@ -353,7 +340,7 @@ Where to find recent results:
          -

          Which Activation and Weights to Choose in Neural Networks

          +

          Which Activation and Weights to Choose in Neural Networks

          1. RELU? ELU?
            2. @@ -376,7 +363,7 @@ Where to find recent results:
            -

            Optimization Methods and Hyperparameters

            +

            Optimization Methods and Hyperparameters

            1. Stochastic gradient descent @@ -401,7 +388,7 @@ Which regularization and hyperparameters? \( L_1 \) or \( L_2 \), soft classifie
              -

              Resampling

              +

              Resampling

              When do we resample? @@ -415,18 +402,17 @@ When do we resample?

              -

              Other courses on Data science and Machine Learning at UiO

              +

              Other courses on Data science and Machine Learning at UiO

              The link here https://www.mn.uio.no/english/research/about/centre-focus/innovation/data-science/studies/ gives an excellent overview of courses on Machine learning at UiO.

              1. STK2100 Machine learning and statistical methods for prediction and classification.
                2. -

              2. IN3050 Introduction to Artificial Intelligence and Machine Learning. Introductory course in machine learning and AI with an algorithmic approach.
                3. +

              3. IN3050/IN4050 Introduction to Artificial Intelligence and Machine Learning. Introductory course in machine learning and AI with an algorithmic approach.

              4. STK-INF3000/4000 Selected Topics in Data Science. The course provides insight into selected contemporary relevant topics within Data Science.

              5. IN4080 Natural Language Processing. Probabilistic and machine learning techniques applied to natural language processing.

              6. STK-IN4300 – Statistical learning methods in Data Science. An advanced introduction to statistical and machine learning. For students with a good mathematics and statistics background.
                7. -

              7. INF4490 Biologically Inspired Computing. An introduction to self-adapting methods also called artificial intelligence or machine learning.

              8. IN-STK5000 Adaptive Methods for Data-Based Decision Making. Methods for adaptive collection and processing of data based on machine learning techniques.

              9. IN5400/INF5860 – Machine Learning for Image Analysis. An introduction to deep learning with particular emphasis on applications within Image analysis, but useful for other application areas too.

              10. TEK5040 – Dyp læring for autonome systemer. The course addresses advanced algorithms and architectures for deep learning with neural networks. The course provides an introduction to how deep-learning techniques can be used in the construction of key parts of advanced autonomous systems that exist in physical environments and cyber environments.
                11. @@ -435,7 +421,7 @@ The link here Additional courses of interest +

                Additional courses of interest

                1. STK4051 Computational Statistics
                  2. @@ -445,7 +431,290 @@ The link here Best wishes to you all and thanks so much for your heroic efforts this semester +

                  Hot Topics Now

                  + +
                    +

                  1. Boosting techniques and complex neural networks
                    2. +

                  2. Adversarial examples
                    3. +

                  3. Zero shot learning
                    4. +

                  4. Transfer learning
                    5. +

                  5. Model interpretability
                    6. +
                  +

                  + +Based on multi-layer nonlinear neural networks, deep learning can +learn directly from raw data, automatically extract and abstract +features from layer to layer, and then achieve the goal of regression, +classification, or ranking. Deep learning has made breakthroughs in +computer vision, speech processing and natural language, and reached +or even surpassed human level. The success of deep learning is mainly +due to the three factors: big data, big model, and big computing. + +

                  +In the past few decades, many different architectures of deep neural +networks have been proposed, such as + +

                    +

                  1. Convolutional neural networks, which are mostly used in image and video data processing, and have also been applied to sequential data such as text processing;
                    2. +

                  2. Recurrent neural networks, which can process sequential data of variable length and have been widely used in natural language understanding and speech processing;
                    3. +

                  3. Encoder-decoder framework, which is mostly used for image or sequence generation, such as machine translation, text summarization, and image captioning.
                    4. +
                  +
              + + +
              +

              Reinforcement Learning

              + +

              +Reinforcement learning is a sub-area of machine learning. It studies +how agents take actions based on trial and error, so as to maximize +some notion of cumulative reward in a dynamic system or +environment. Due to its generality, the problem has also been studied +in many other disciplines, such as game theory, control theory, +operations research, information theory, multi-agent systems, swarm +intelligence, statistics, and genetic algorithms. + +

              +In March 2016, AlphaGo, a computer program that plays the board game +Go, beat Lee Sedol in a five-game match. This was the first time a +computer Go program had beaten a 9-dan (highest rank) professional +without handicaps. AlphaGo is based on deep convolutional neural +networks and reinforcement learning. AlphaGo’s victory was a major +milestone in artificial intelligence and it has also made +reinforcement learning a hot research area in the field of machine +learning. +

              + + +
              +

              Transfer learning

              + +

              +The goal of transfer learning is to transfer the model or knowledge +obtained from a source task to the target task, in order to resolve +the issues of insufficient training data in the target task. The +rationality of doing so lies in that usually the source and target +tasks have inter-correlations, and therefore either the features, +samples, or models in the source task might provide useful information +for us to better solve the target task. Transfer learning is a hot +research topic in recent years, with many problems still waiting to be +solved in this space. +

              + + +
              +

              Adversarial learning

              + +

              +The conventional deep generative model has a potential problem: the +model tends to generate extreme instances to maximize the +probabilistic likelihood, which will hurt its performance. Adversarial +learning utilizes the adversarial behaviors (e.g., generating +adversarial instances or training an adversarial model) to enhance the +robustness of the model and improve the quality of the generated +data. In recent years, one of the most promising unsupervised learning +technologies, generative adversarial networks (GAN), has already been +successfully applied to image, speech, and text. +

              + + +
              +

              Dual learning

              + +

              +Dual learning is a new learning paradigm, the basic idea of which is +to use the primal-dual structure between machine learning tasks to +obtain effective feedback/regularization, and guide and strengthen the +learning process, thus reducing the requirement of large-scale labeled +data for deep learning. The idea of dual learning has been applied to +many problems in machine learning, including machine translation, +image style conversion, question answering and generation, image +classification and generation, text classification and generation, +image-to-text, and text-to-image. +

              + + +
              +

              Distributed machine learning

              + +

              +Distributed computation will speed up machine learning algorithms, +significantly improve their efficiency, and thus enlarge their +application. When distributed meets machine learning, more than just +implementing the machine learning algorithms in parallel is required. +

              + + +
              +

              Meta learning

              + +

              +Meta learning is an emerging research direction in machine +learning. Roughly speaking, meta learning concerns learning how to +learn, and focuses on the understanding and adaptation of the learning +itself, instead of just completing a specific learning task. That is, +a meta learner needs to be able to evaluate its own learning methods +and adjust its own learning methods according to specific learning +tasks. +

              + + +
              +

              The Challenges Facing Machine Learning

              + +

              +While there has been much progress in machine learning, there are also challenges. + +

              +For example, the mainstream machine learning technologies are +black-box approaches, making us concerned about their potential +risks. To tackle this challenge, we may want to make machine learning +more explainable and controllable. As another example, the +computational complexity of machine learning algorithms is usually +very high and we may want to invent lightweight algorithms or +implementations. Furthermore, in many domains such as physics, +chemistry, biology, and social sciences, people usually seek elegantly +simple equations (e.g., the Schrödinger equation) to uncover the +underlying laws behind various phenomena. In the field of machine +learning, can we reveal simple laws instead of designing more complex +models for data fitting? Although there are many challenges, we are +still very optimistic about the future of machine learning. As we look +forward to the future, here are what we think the research hotspots in +the next ten years will be. +

              + + +
              +

              Explainable machine learning

              + +

              +Machine learning, especially deep learning, evolves rapidly. The +ability gap between machine and human on many complex cognitive tasks +becomes narrower and narrower. However, we are still in the very early +stage in terms of explaining why those effective models work and how +they work. + +

              +What is missing: the gap between correlation and causation Most +machine learning techniques, especially the statistical ones, depend +highly on data correlation to make predictions and analyses. In +contrast, rational humans tend to reply on clear and trustworthy +causality relations obtained via logical reasoning on real and clear +facts. It is one of the core goals of explainable machine learning to +transition from solving problems by data correlation to solving +problems by logical reasoning. +

              + + +
              +

              Quantum machine learning

              + +

              +Quantum machine learning is an emerging interdisciplinary research +area at the intersection of quantum computing and machine learning. + +

              +Quantum computers use effects such as quantum coherence and quantum +entanglement to process information, which is fundamentally different +from classical computers. Quantum algorithms have surpassed the best +classical algorithms in several problems (e.g., searching for an +unsorted database, inverting a sparse matrix), which we call quantum +acceleration. + +

              +When quantum computing meets machine learning, it can be a mutually +beneficial and reinforcing process, as it allows us to take advantage +of quantum computing to improve the performance of classical machine +learning algorithms. In addition, we can also use the machine learning +algorithms (on classic computers) to analyze and improve quantum +computing systems. +

              + + +
              +

              Quantum machine learning algorithms based on linear algebra

              + +

              +Many quantum machine learning algorithms are based on variants of +quantum algorithms for solving linear equations, which can efficiently +solve N-variable linear equations with complexity of O(log2 N) under +certain conditions. The quantum matrix inversion algorithm can +accelerate many machine learning methods, such as least square linear +regression, least square version of support vector machine, Gaussian +process, and more. The training of these algorithms can be simplified +to solve linear equations. The key bottleneck of this type of quantum +machine learning algorithms is data input—that is, how to initialize +the quantum system with the entire data set. Although efficient +data-input algorithms exist for certain situations, how to efficiently +input data into a quantum system is as yet unknown for most cases. +

              + + +
              +

              Quantum reinforcement learning

              + +

              +In quantum reinforcement learning, a quantum agent interacts with the +classical environment to obtain rewards from the environment, so as to +adjust and improve its behavioral strategies. In some cases, it +achieves quantum acceleration by the quantum processing capabilities +of the agent or the possibility of exploring the environment through +quantum superposition. Such algorithms have been proposed in +superconducting circuits and systems of trapped ions. +

              + + +
              +

              Quantum deep learning

              + +

              +Dedicated quantum information processors, such as quantum annealers +and programmable photonic circuits, are well suited for building deep +quantum networks. The simplest deep quantum network is the Boltzmann +machine. The classical Boltzmann machine consists of bits with tunable +interactions and is trained by adjusting the interaction of these bits +so that the distribution of its expression conforms to the statistics +of the data. To quantize the Boltzmann machine, the neural network can +simply be represented as a set of interacting quantum spins that +correspond to an adjustable Ising model. Then, by initializing the +input neurons in the Boltzmann machine to a fixed state and allowing +the system to heat up, we can read out the output qubits to get the +result. +

              + + +
              +

              Social machine learning

              + +

              +Machine learning aims to imitate how humans +learn. While we have developed successful machine learning algorithms, +until now we have ignored one important fact: humans are social. Each +of us is one part of the total society and it is difficult for us to +live, learn, and improve ourselves, alone and isolated. Therefore, we +should design machines with social properties. Can we let machines +evolve by imitating human society so as to achieve more effective, +intelligent, interpretable “social machine learning”? + +

              +And much more. +

              + + +
              +

              The last words?

              + +

              +Early computer scientist Alan Kay said, The best way to predict the +future is to create it. Therefore, all machine learning +practitioners, whether scholars or engineers, professors or students, +need to work together to advance these important research +topics. Together, we will not just predict the future, but create it. +

              + + +
              +

              Best wishes to you all and thanks so much for your heroic efforts this semester





              diff --git a/doc/pub/summary/html/summary-solarized.html b/doc/pub/summary/html/summary-solarized.html index 8e005268d..dc99907f6 100644 --- a/doc/pub/summary/html/summary-solarized.html +++ b/doc/pub/summary/html/summary-solarized.html @@ -52,26 +52,43 @@ div { text-align: justify; text-justify: inter-word; } '___sec5'), ('Perspective on Machine Learning', 2, None, '___sec6'), ('Machine Learning Research', 2, None, '___sec7'), - ('Hot Topics Now', 2, None, '___sec8'), - ('Starting your Machine Learning Project', 2, None, '___sec9'), - ('Choose a Model and Algorithm', 2, None, '___sec10'), - ('Preparing Your Data', 2, None, '___sec11'), + ('Starting your Machine Learning Project', 2, None, '___sec8'), + ('Choose a Model and Algorithm', 2, None, '___sec9'), + ('Preparing Your Data', 2, None, '___sec10'), ('Which Activation and Weights to Choose in Neural Networks', 2, None, - '___sec12'), - ('Optimization Methods and Hyperparameters', 2, None, '___sec13'), - ('Resampling', 2, None, '___sec14'), + '___sec11'), + ('Optimization Methods and Hyperparameters', 2, None, '___sec12'), + ('Resampling', 2, None, '___sec13'), ('Other courses on Data science and Machine Learning at UiO', 2, None, - '___sec15'), - ('Additional courses of interest', 2, None, '___sec16'), + '___sec14'), + ('Additional courses of interest', 2, None, '___sec15'), + ('Hot Topics Now', 2, None, '___sec16'), + ('Reinforcement Learning', 2, None, '___sec17'), + ('Transfer learning', 2, None, '___sec18'), + ('Adversarial learning', 2, None, '___sec19'), + ('Dual learning', 2, None, '___sec20'), + ('Distributed machine learning', 2, None, '___sec21'), + ('Meta learning', 2, None, '___sec22'), + ('The Challenges Facing Machine Learning', 2, None, '___sec23'), + ('Explainable machine learning', 2, None, '___sec24'), + ('Quantum machine learning', 2, None, '___sec25'), + ('Quantum machine learning algorithms based on linear algebra', + 2, + None, + '___sec26'), + ('Quantum reinforcement learning', 2, None, '___sec27'), + ('Quantum deep learning', 2, None, '___sec28'), + ('Social machine learning', 2, None, '___sec29'), + ('The last words?', 2, None, '___sec30'), ('Best wishes to you all and thanks so much for your heroic ' 'efforts this semester', 2, None, - '___sec17')]} + '___sec31')]} end of tocinfo --> @@ -239,19 +256,7 @@ Where to find recent results:









              -

              Hot Topics Now

              - -
                -
              1. Boosting techniques and complex neural networks
                2. -
              2. Adversarial examples
                3. -
              3. Zero shot learning
                4. -
              4. Transfer learning
                5. -
              5. Model interpretability
                6. -
              - -









              - -

              Starting your Machine Learning Project

              +

              Starting your Machine Learning Project

              1. Identify problem type: classification, generation, regression
                2. @@ -263,7 +268,7 @@ Where to find recent results:









                -

                Choose a Model and Algorithm

                +

                Choose a Model and Algorithm

                1. Supervised?
                  2. @@ -273,7 +278,7 @@ Where to find recent results:









                  -

                  Preparing Your Data

                  +

                  Preparing Your Data

                  1. Shuffle your data
                    2. @@ -301,7 +306,7 @@ Where to find recent results:









                    -

                    Which Activation and Weights to Choose in Neural Networks

                    +

                    Which Activation and Weights to Choose in Neural Networks

                    1. RELU? ELU?
                      2. @@ -322,7 +327,7 @@ Where to find recent results:









                      -

                      Optimization Methods and Hyperparameters

                      +

                      Optimization Methods and Hyperparameters

                      1. Stochastic gradient descent @@ -345,7 +350,7 @@ Which regularization and hyperparameters? \( L_1 \) or \( L_2 \), soft classifie











                        -

                        Resampling

                        +

                        Resampling

                        When do we resample? @@ -358,18 +363,17 @@ When do we resample?









                        -

                        Other courses on Data science and Machine Learning at UiO

                        +

                        Other courses on Data science and Machine Learning at UiO

                        The link here https://www.mn.uio.no/english/research/about/centre-focus/innovation/data-science/studies/ gives an excellent overview of courses on Machine learning at UiO.

                        1. STK2100 Machine learning and statistical methods for prediction and classification.
                          2. -
                        2. IN3050 Introduction to Artificial Intelligence and Machine Learning. Introductory course in machine learning and AI with an algorithmic approach.
                          3. +
                        3. IN3050/IN4050 Introduction to Artificial Intelligence and Machine Learning. Introductory course in machine learning and AI with an algorithmic approach.
                        4. STK-INF3000/4000 Selected Topics in Data Science. The course provides insight into selected contemporary relevant topics within Data Science.
                        5. IN4080 Natural Language Processing. Probabilistic and machine learning techniques applied to natural language processing.
                        6. STK-IN4300 – Statistical learning methods in Data Science. An advanced introduction to statistical and machine learning. For students with a good mathematics and statistics background.
                          7. -
                        7. INF4490 Biologically Inspired Computing. An introduction to self-adapting methods also called artificial intelligence or machine learning.
                        8. IN-STK5000 Adaptive Methods for Data-Based Decision Making. Methods for adaptive collection and processing of data based on machine learning techniques.
                        9. IN5400/INF5860 – Machine Learning for Image Analysis. An introduction to deep learning with particular emphasis on applications within Image analysis, but useful for other application areas too.
                        10. TEK5040 – Dyp læring for autonome systemer. The course addresses advanced algorithms and architectures for deep learning with neural networks. The course provides an introduction to how deep-learning techniques can be used in the construction of key parts of advanced autonomous systems that exist in physical environments and cyber environments.
                          11. @@ -377,7 +381,7 @@ The link here Additional courses of interest +

                          Additional courses of interest

                          1. STK4051 Computational Statistics
                            2. @@ -386,13 +390,292 @@ The link here Best wishes to you all and thanks so much for your heroic efforts this semester +

                            Hot Topics Now

                            + +
                              +
                            1. Boosting techniques and complex neural networks
                              2. +
                            2. Adversarial examples
                              3. +
                            3. Zero shot learning
                              4. +
                            4. Transfer learning
                              5. +
                            5. Model interpretability
                              6. +
                            + +Based on multi-layer nonlinear neural networks, deep learning can +learn directly from raw data, automatically extract and abstract +features from layer to layer, and then achieve the goal of regression, +classification, or ranking. Deep learning has made breakthroughs in +computer vision, speech processing and natural language, and reached +or even surpassed human level. The success of deep learning is mainly +due to the three factors: big data, big model, and big computing. + +

                            +In the past few decades, many different architectures of deep neural +networks have been proposed, such as + +

                              +
                            1. Convolutional neural networks, which are mostly used in image and video data processing, and have also been applied to sequential data such as text processing;
                              2. +
                            2. Recurrent neural networks, which can process sequential data of variable length and have been widely used in natural language understanding and speech processing;
                              3. +
                            3. Encoder-decoder framework, which is mostly used for image or sequence generation, such as machine translation, text summarization, and image captioning.
                              4. +
                            + +









                            + +

                            Reinforcement Learning

                            + +

                            +Reinforcement learning is a sub-area of machine learning. It studies +how agents take actions based on trial and error, so as to maximize +some notion of cumulative reward in a dynamic system or +environment. Due to its generality, the problem has also been studied +in many other disciplines, such as game theory, control theory, +operations research, information theory, multi-agent systems, swarm +intelligence, statistics, and genetic algorithms. + +

                            +In March 2016, AlphaGo, a computer program that plays the board game +Go, beat Lee Sedol in a five-game match. This was the first time a +computer Go program had beaten a 9-dan (highest rank) professional +without handicaps. AlphaGo is based on deep convolutional neural +networks and reinforcement learning. AlphaGo’s victory was a major +milestone in artificial intelligence and it has also made +reinforcement learning a hot research area in the field of machine +learning. + +

                            +









                            + +

                            Transfer learning

                            + +

                            +The goal of transfer learning is to transfer the model or knowledge +obtained from a source task to the target task, in order to resolve +the issues of insufficient training data in the target task. The +rationality of doing so lies in that usually the source and target +tasks have inter-correlations, and therefore either the features, +samples, or models in the source task might provide useful information +for us to better solve the target task. Transfer learning is a hot +research topic in recent years, with many problems still waiting to be +solved in this space. + +

                            +









                            + +

                            Adversarial learning

                            + +

                            +The conventional deep generative model has a potential problem: the +model tends to generate extreme instances to maximize the +probabilistic likelihood, which will hurt its performance. Adversarial +learning utilizes the adversarial behaviors (e.g., generating +adversarial instances or training an adversarial model) to enhance the +robustness of the model and improve the quality of the generated +data. In recent years, one of the most promising unsupervised learning +technologies, generative adversarial networks (GAN), has already been +successfully applied to image, speech, and text. + +

                            +









                            + +

                            Dual learning

                            + +

                            +Dual learning is a new learning paradigm, the basic idea of which is +to use the primal-dual structure between machine learning tasks to +obtain effective feedback/regularization, and guide and strengthen the +learning process, thus reducing the requirement of large-scale labeled +data for deep learning. The idea of dual learning has been applied to +many problems in machine learning, including machine translation, +image style conversion, question answering and generation, image +classification and generation, text classification and generation, +image-to-text, and text-to-image. + +

                            +









                            + +

                            Distributed machine learning

                            + +

                            +Distributed computation will speed up machine learning algorithms, +significantly improve their efficiency, and thus enlarge their +application. When distributed meets machine learning, more than just +implementing the machine learning algorithms in parallel is required. + +

                            +









                            + +

                            Meta learning

                            + +

                            +Meta learning is an emerging research direction in machine +learning. Roughly speaking, meta learning concerns learning how to +learn, and focuses on the understanding and adaptation of the learning +itself, instead of just completing a specific learning task. That is, +a meta learner needs to be able to evaluate its own learning methods +and adjust its own learning methods according to specific learning +tasks. + +

                            +









                            + +

                            The Challenges Facing Machine Learning

                            + +

                            +While there has been much progress in machine learning, there are also challenges. + +

                            +For example, the mainstream machine learning technologies are +black-box approaches, making us concerned about their potential +risks. To tackle this challenge, we may want to make machine learning +more explainable and controllable. As another example, the +computational complexity of machine learning algorithms is usually +very high and we may want to invent lightweight algorithms or +implementations. Furthermore, in many domains such as physics, +chemistry, biology, and social sciences, people usually seek elegantly +simple equations (e.g., the Schrödinger equation) to uncover the +underlying laws behind various phenomena. In the field of machine +learning, can we reveal simple laws instead of designing more complex +models for data fitting? Although there are many challenges, we are +still very optimistic about the future of machine learning. As we look +forward to the future, here are what we think the research hotspots in +the next ten years will be. + +

                            +









                            + +

                            Explainable machine learning

                            + +

                            +Machine learning, especially deep learning, evolves rapidly. The +ability gap between machine and human on many complex cognitive tasks +becomes narrower and narrower. However, we are still in the very early +stage in terms of explaining why those effective models work and how +they work. + +

                            +What is missing: the gap between correlation and causation Most +machine learning techniques, especially the statistical ones, depend +highly on data correlation to make predictions and analyses. In +contrast, rational humans tend to reply on clear and trustworthy +causality relations obtained via logical reasoning on real and clear +facts. It is one of the core goals of explainable machine learning to +transition from solving problems by data correlation to solving +problems by logical reasoning. + +

                            +









                            + +

                            Quantum machine learning

                            + +

                            +Quantum machine learning is an emerging interdisciplinary research +area at the intersection of quantum computing and machine learning. + +

                            +Quantum computers use effects such as quantum coherence and quantum +entanglement to process information, which is fundamentally different +from classical computers. Quantum algorithms have surpassed the best +classical algorithms in several problems (e.g., searching for an +unsorted database, inverting a sparse matrix), which we call quantum +acceleration. + +

                            +When quantum computing meets machine learning, it can be a mutually +beneficial and reinforcing process, as it allows us to take advantage +of quantum computing to improve the performance of classical machine +learning algorithms. In addition, we can also use the machine learning +algorithms (on classic computers) to analyze and improve quantum +computing systems. + +

                            +









                            + +

                            Quantum machine learning algorithms based on linear algebra

                            + +

                            +Many quantum machine learning algorithms are based on variants of +quantum algorithms for solving linear equations, which can efficiently +solve N-variable linear equations with complexity of O(log2 N) under +certain conditions. The quantum matrix inversion algorithm can +accelerate many machine learning methods, such as least square linear +regression, least square version of support vector machine, Gaussian +process, and more. The training of these algorithms can be simplified +to solve linear equations. The key bottleneck of this type of quantum +machine learning algorithms is data input—that is, how to initialize +the quantum system with the entire data set. Although efficient +data-input algorithms exist for certain situations, how to efficiently +input data into a quantum system is as yet unknown for most cases. + +

                            +









                            + +

                            Quantum reinforcement learning

                            + +

                            +In quantum reinforcement learning, a quantum agent interacts with the +classical environment to obtain rewards from the environment, so as to +adjust and improve its behavioral strategies. In some cases, it +achieves quantum acceleration by the quantum processing capabilities +of the agent or the possibility of exploring the environment through +quantum superposition. Such algorithms have been proposed in +superconducting circuits and systems of trapped ions. + +

                            +









                            + +

                            Quantum deep learning

                            + +

                            +Dedicated quantum information processors, such as quantum annealers +and programmable photonic circuits, are well suited for building deep +quantum networks. The simplest deep quantum network is the Boltzmann +machine. The classical Boltzmann machine consists of bits with tunable +interactions and is trained by adjusting the interaction of these bits +so that the distribution of its expression conforms to the statistics +of the data. To quantize the Boltzmann machine, the neural network can +simply be represented as a set of interacting quantum spins that +correspond to an adjustable Ising model. Then, by initializing the +input neurons in the Boltzmann machine to a fixed state and allowing +the system to heat up, we can read out the output qubits to get the +result. + +

                            +









                            + +

                            Social machine learning

                            + +

                            +Machine learning aims to imitate how humans +learn. While we have developed successful machine learning algorithms, +until now we have ignored one important fact: humans are social. Each +of us is one part of the total society and it is difficult for us to +live, learn, and improve ourselves, alone and isolated. Therefore, we +should design machines with social properties. Can we let machines +evolve by imitating human society so as to achieve more effective, +intelligent, interpretable “social machine learning”? + +

                            +And much more. + +

                            +









                            + +

                            The last words?

                            + +

                            +Early computer scientist Alan Kay said, The best way to predict the +future is to create it. Therefore, all machine learning +practitioners, whether scholars or engineers, professors or students, +need to work together to advance these important research +topics. Together, we will not just predict the future, but create it. + +

                            +









                            + +

                            Best wishes to you all and thanks so much for your heroic efforts this semester





                            -

                            - diff --git a/doc/pub/summary/html/summary.html b/doc/pub/summary/html/summary.html index 89656495a..e2a75ab16 100644 --- a/doc/pub/summary/html/summary.html +++ b/doc/pub/summary/html/summary.html @@ -57,26 +57,43 @@ div { text-align: justify; text-justify: inter-word; } '___sec5'), ('Perspective on Machine Learning', 2, None, '___sec6'), ('Machine Learning Research', 2, None, '___sec7'), - ('Hot Topics Now', 2, None, '___sec8'), - ('Starting your Machine Learning Project', 2, None, '___sec9'), - ('Choose a Model and Algorithm', 2, None, '___sec10'), - ('Preparing Your Data', 2, None, '___sec11'), + ('Starting your Machine Learning Project', 2, None, '___sec8'), + ('Choose a Model and Algorithm', 2, None, '___sec9'), + ('Preparing Your Data', 2, None, '___sec10'), ('Which Activation and Weights to Choose in Neural Networks', 2, None, - '___sec12'), - ('Optimization Methods and Hyperparameters', 2, None, '___sec13'), - ('Resampling', 2, None, '___sec14'), + '___sec11'), + ('Optimization Methods and Hyperparameters', 2, None, '___sec12'), + ('Resampling', 2, None, '___sec13'), ('Other courses on Data science and Machine Learning at UiO', 2, None, - '___sec15'), - ('Additional courses of interest', 2, None, '___sec16'), + '___sec14'), + ('Additional courses of interest', 2, None, '___sec15'), + ('Hot Topics Now', 2, None, '___sec16'), + ('Reinforcement Learning', 2, None, '___sec17'), + ('Transfer learning', 2, None, '___sec18'), + ('Adversarial learning', 2, None, '___sec19'), + ('Dual learning', 2, None, '___sec20'), + ('Distributed machine learning', 2, None, '___sec21'), + ('Meta learning', 2, None, '___sec22'), + ('The Challenges Facing Machine Learning', 2, None, '___sec23'), + ('Explainable machine learning', 2, None, '___sec24'), + ('Quantum machine learning', 2, None, '___sec25'), + ('Quantum machine learning algorithms based on linear algebra', + 2, + None, + '___sec26'), + ('Quantum reinforcement learning', 2, None, '___sec27'), + ('Quantum deep learning', 2, None, '___sec28'), + ('Social machine learning', 2, None, '___sec29'), + ('The last words?', 2, None, '___sec30'), ('Best wishes to you all and thanks so much for your heroic ' 'efforts this semester', 2, None, - '___sec17')]} + '___sec31')]} end of tocinfo --> @@ -244,19 +261,7 @@ Where to find recent results:









                            -

                            Hot Topics Now

                            - -
                              -
                            1. Boosting techniques and complex neural networks
                              2. -
                            2. Adversarial examples
                              3. -
                            3. Zero shot learning
                              4. -
                            4. Transfer learning
                              5. -
                            5. Model interpretability
                              6. -
                            - -









                            - -

                            Starting your Machine Learning Project

                            +

                            Starting your Machine Learning Project

                            1. Identify problem type: classification, generation, regression
                              2. @@ -268,7 +273,7 @@ Where to find recent results:









                              -

                              Choose a Model and Algorithm

                              +

                              Choose a Model and Algorithm

                              1. Supervised?
                                2. @@ -278,7 +283,7 @@ Where to find recent results:









                                -

                                Preparing Your Data

                                +

                                Preparing Your Data

                                1. Shuffle your data
                                  2. @@ -306,7 +311,7 @@ Where to find recent results:









                                  -

                                  Which Activation and Weights to Choose in Neural Networks

                                  +

                                  Which Activation and Weights to Choose in Neural Networks

                                  1. RELU? ELU?
                                    2. @@ -327,7 +332,7 @@ Where to find recent results:









                                    -

                                    Optimization Methods and Hyperparameters

                                    +

                                    Optimization Methods and Hyperparameters

                                    1. Stochastic gradient descent @@ -350,7 +355,7 @@ Which regularization and hyperparameters? \( L_1 \) or \( L_2 \), soft classifie











                                      -

                                      Resampling

                                      +

                                      Resampling

                                      When do we resample? @@ -363,18 +368,17 @@ When do we resample?









                                      -

                                      Other courses on Data science and Machine Learning at UiO

                                      +

                                      Other courses on Data science and Machine Learning at UiO

                                      The link here https://www.mn.uio.no/english/research/about/centre-focus/innovation/data-science/studies/ gives an excellent overview of courses on Machine learning at UiO.

                                      1. STK2100 Machine learning and statistical methods for prediction and classification.
                                        2. -
                                      2. IN3050 Introduction to Artificial Intelligence and Machine Learning. Introductory course in machine learning and AI with an algorithmic approach.
                                        3. +
                                      3. IN3050/IN4050 Introduction to Artificial Intelligence and Machine Learning. Introductory course in machine learning and AI with an algorithmic approach.
                                      4. STK-INF3000/4000 Selected Topics in Data Science. The course provides insight into selected contemporary relevant topics within Data Science.
                                      5. IN4080 Natural Language Processing. Probabilistic and machine learning techniques applied to natural language processing.
                                      6. STK-IN4300 – Statistical learning methods in Data Science. An advanced introduction to statistical and machine learning. For students with a good mathematics and statistics background.
                                        7. -
                                      7. INF4490 Biologically Inspired Computing. An introduction to self-adapting methods also called artificial intelligence or machine learning.
                                      8. IN-STK5000 Adaptive Methods for Data-Based Decision Making. Methods for adaptive collection and processing of data based on machine learning techniques.
                                      9. IN5400/INF5860 – Machine Learning for Image Analysis. An introduction to deep learning with particular emphasis on applications within Image analysis, but useful for other application areas too.
                                      10. TEK5040 – Dyp læring for autonome systemer. The course addresses advanced algorithms and architectures for deep learning with neural networks. The course provides an introduction to how deep-learning techniques can be used in the construction of key parts of advanced autonomous systems that exist in physical environments and cyber environments.
                                        11. @@ -382,7 +386,7 @@ The link here Additional courses of interest +

                                        Additional courses of interest

                                        1. STK4051 Computational Statistics
                                          2. @@ -391,13 +395,292 @@ The link here Best wishes to you all and thanks so much for your heroic efforts this semester +

                                          Hot Topics Now

                                          + +
                                            +
                                          1. Boosting techniques and complex neural networks
                                            2. +
                                          2. Adversarial examples
                                            3. +
                                          3. Zero shot learning
                                            4. +
                                          4. Transfer learning
                                            5. +
                                          5. Model interpretability
                                            6. +
                                          + +Based on multi-layer nonlinear neural networks, deep learning can +learn directly from raw data, automatically extract and abstract +features from layer to layer, and then achieve the goal of regression, +classification, or ranking. Deep learning has made breakthroughs in +computer vision, speech processing and natural language, and reached +or even surpassed human level. The success of deep learning is mainly +due to the three factors: big data, big model, and big computing. + +

                                          +In the past few decades, many different architectures of deep neural +networks have been proposed, such as + +

                                            +
                                          1. Convolutional neural networks, which are mostly used in image and video data processing, and have also been applied to sequential data such as text processing;
                                            2. +
                                          2. Recurrent neural networks, which can process sequential data of variable length and have been widely used in natural language understanding and speech processing;
                                            3. +
                                          3. Encoder-decoder framework, which is mostly used for image or sequence generation, such as machine translation, text summarization, and image captioning.
                                            4. +
                                          + +









                                          + +

                                          Reinforcement Learning

                                          + +

                                          +Reinforcement learning is a sub-area of machine learning. It studies +how agents take actions based on trial and error, so as to maximize +some notion of cumulative reward in a dynamic system or +environment. Due to its generality, the problem has also been studied +in many other disciplines, such as game theory, control theory, +operations research, information theory, multi-agent systems, swarm +intelligence, statistics, and genetic algorithms. + +

                                          +In March 2016, AlphaGo, a computer program that plays the board game +Go, beat Lee Sedol in a five-game match. This was the first time a +computer Go program had beaten a 9-dan (highest rank) professional +without handicaps. AlphaGo is based on deep convolutional neural +networks and reinforcement learning. AlphaGo’s victory was a major +milestone in artificial intelligence and it has also made +reinforcement learning a hot research area in the field of machine +learning. + +

                                          +









                                          + +

                                          Transfer learning

                                          + +

                                          +The goal of transfer learning is to transfer the model or knowledge +obtained from a source task to the target task, in order to resolve +the issues of insufficient training data in the target task. The +rationality of doing so lies in that usually the source and target +tasks have inter-correlations, and therefore either the features, +samples, or models in the source task might provide useful information +for us to better solve the target task. Transfer learning is a hot +research topic in recent years, with many problems still waiting to be +solved in this space. + +

                                          +









                                          + +

                                          Adversarial learning

                                          + +

                                          +The conventional deep generative model has a potential problem: the +model tends to generate extreme instances to maximize the +probabilistic likelihood, which will hurt its performance. Adversarial +learning utilizes the adversarial behaviors (e.g., generating +adversarial instances or training an adversarial model) to enhance the +robustness of the model and improve the quality of the generated +data. In recent years, one of the most promising unsupervised learning +technologies, generative adversarial networks (GAN), has already been +successfully applied to image, speech, and text. + +

                                          +









                                          + +

                                          Dual learning

                                          + +

                                          +Dual learning is a new learning paradigm, the basic idea of which is +to use the primal-dual structure between machine learning tasks to +obtain effective feedback/regularization, and guide and strengthen the +learning process, thus reducing the requirement of large-scale labeled +data for deep learning. The idea of dual learning has been applied to +many problems in machine learning, including machine translation, +image style conversion, question answering and generation, image +classification and generation, text classification and generation, +image-to-text, and text-to-image. + +

                                          +









                                          + +

                                          Distributed machine learning

                                          + +

                                          +Distributed computation will speed up machine learning algorithms, +significantly improve their efficiency, and thus enlarge their +application. When distributed meets machine learning, more than just +implementing the machine learning algorithms in parallel is required. + +

                                          +









                                          + +

                                          Meta learning

                                          + +

                                          +Meta learning is an emerging research direction in machine +learning. Roughly speaking, meta learning concerns learning how to +learn, and focuses on the understanding and adaptation of the learning +itself, instead of just completing a specific learning task. That is, +a meta learner needs to be able to evaluate its own learning methods +and adjust its own learning methods according to specific learning +tasks. + +

                                          +









                                          + +

                                          The Challenges Facing Machine Learning

                                          + +

                                          +While there has been much progress in machine learning, there are also challenges. + +

                                          +For example, the mainstream machine learning technologies are +black-box approaches, making us concerned about their potential +risks. To tackle this challenge, we may want to make machine learning +more explainable and controllable. As another example, the +computational complexity of machine learning algorithms is usually +very high and we may want to invent lightweight algorithms or +implementations. Furthermore, in many domains such as physics, +chemistry, biology, and social sciences, people usually seek elegantly +simple equations (e.g., the Schrödinger equation) to uncover the +underlying laws behind various phenomena. In the field of machine +learning, can we reveal simple laws instead of designing more complex +models for data fitting? Although there are many challenges, we are +still very optimistic about the future of machine learning. As we look +forward to the future, here are what we think the research hotspots in +the next ten years will be. + +

                                          +









                                          + +

                                          Explainable machine learning

                                          + +

                                          +Machine learning, especially deep learning, evolves rapidly. The +ability gap between machine and human on many complex cognitive tasks +becomes narrower and narrower. However, we are still in the very early +stage in terms of explaining why those effective models work and how +they work. + +

                                          +What is missing: the gap between correlation and causation Most +machine learning techniques, especially the statistical ones, depend +highly on data correlation to make predictions and analyses. In +contrast, rational humans tend to reply on clear and trustworthy +causality relations obtained via logical reasoning on real and clear +facts. It is one of the core goals of explainable machine learning to +transition from solving problems by data correlation to solving +problems by logical reasoning. + +

                                          +









                                          + +

                                          Quantum machine learning

                                          + +

                                          +Quantum machine learning is an emerging interdisciplinary research +area at the intersection of quantum computing and machine learning. + +

                                          +Quantum computers use effects such as quantum coherence and quantum +entanglement to process information, which is fundamentally different +from classical computers. Quantum algorithms have surpassed the best +classical algorithms in several problems (e.g., searching for an +unsorted database, inverting a sparse matrix), which we call quantum +acceleration. + +

                                          +When quantum computing meets machine learning, it can be a mutually +beneficial and reinforcing process, as it allows us to take advantage +of quantum computing to improve the performance of classical machine +learning algorithms. In addition, we can also use the machine learning +algorithms (on classic computers) to analyze and improve quantum +computing systems. + +

                                          +









                                          + +

                                          Quantum machine learning algorithms based on linear algebra

                                          + +

                                          +Many quantum machine learning algorithms are based on variants of +quantum algorithms for solving linear equations, which can efficiently +solve N-variable linear equations with complexity of O(log2 N) under +certain conditions. The quantum matrix inversion algorithm can +accelerate many machine learning methods, such as least square linear +regression, least square version of support vector machine, Gaussian +process, and more. The training of these algorithms can be simplified +to solve linear equations. The key bottleneck of this type of quantum +machine learning algorithms is data input—that is, how to initialize +the quantum system with the entire data set. Although efficient +data-input algorithms exist for certain situations, how to efficiently +input data into a quantum system is as yet unknown for most cases. + +

                                          +









                                          + +

                                          Quantum reinforcement learning

                                          + +

                                          +In quantum reinforcement learning, a quantum agent interacts with the +classical environment to obtain rewards from the environment, so as to +adjust and improve its behavioral strategies. In some cases, it +achieves quantum acceleration by the quantum processing capabilities +of the agent or the possibility of exploring the environment through +quantum superposition. Such algorithms have been proposed in +superconducting circuits and systems of trapped ions. + +

                                          +









                                          + +

                                          Quantum deep learning

                                          + +

                                          +Dedicated quantum information processors, such as quantum annealers +and programmable photonic circuits, are well suited for building deep +quantum networks. The simplest deep quantum network is the Boltzmann +machine. The classical Boltzmann machine consists of bits with tunable +interactions and is trained by adjusting the interaction of these bits +so that the distribution of its expression conforms to the statistics +of the data. To quantize the Boltzmann machine, the neural network can +simply be represented as a set of interacting quantum spins that +correspond to an adjustable Ising model. Then, by initializing the +input neurons in the Boltzmann machine to a fixed state and allowing +the system to heat up, we can read out the output qubits to get the +result. + +

                                          +









                                          + +

                                          Social machine learning

                                          + +

                                          +Machine learning aims to imitate how humans +learn. While we have developed successful machine learning algorithms, +until now we have ignored one important fact: humans are social. Each +of us is one part of the total society and it is difficult for us to +live, learn, and improve ourselves, alone and isolated. Therefore, we +should design machines with social properties. Can we let machines +evolve by imitating human society so as to achieve more effective, +intelligent, interpretable “social machine learning”? + +

                                          +And much more. + +

                                          +









                                          + +

                                          The last words?

                                          + +

                                          +Early computer scientist Alan Kay said, The best way to predict the +future is to create it. Therefore, all machine learning +practitioners, whether scholars or engineers, professors or students, +need to work together to advance these important research +topics. Together, we will not just predict the future, but create it. + +

                                          +









                                          + +

                                          Best wishes to you all and thanks so much for your heroic efforts this semester





                                          -

                                          - diff --git a/doc/pub/summary/ipynb/ipynb-summary-src.tar.gz b/doc/pub/summary/ipynb/ipynb-summary-src.tar.gz index e90ff59919160f4351e898c53ec559e2e0c4bcf5..dbc79daf1e9c88dccecacf95d3750822d2ce8cb2 100644 GIT binary patch delta 148 zcmWN_xdFid06((U&^0x3&H zLa9nIHHoAyu{5M9EonBPSkfDrZEQw5HDl?hOLYA_URMxVQt?Xnk2RX_~&T^6K Im)rgK2W4|2#Q*>R diff --git a/doc/pub/summary/ipynb/summary.ipynb b/doc/pub/summary/ipynb/summary.ipynb index 6f613f755..3cea2fa8d 100644 --- a/doc/pub/summary/ipynb/summary.ipynb +++ b/doc/pub/summary/ipynb/summary.ipynb @@ -146,18 +146,6 @@ "\n", "5. [Journal of Machine Learning Research](http://www.jmlr.org/papers/v19/) \n", "\n", - "## Hot Topics Now\n", - "\n", - "1. Boosting techniques and complex neural networks\n", - "\n", - "2. [Adversarial examples](https://medium.com/@ml.at.berkeley/trickingneural-networks-create-your-own-adversarial-examples-a61eb7620fd8)\n", - "\n", - "3. [Zero shot learning](https://arxiv.org/pdf/1707.00600)\n", - "\n", - "4. Transfer learning\n", - "\n", - "5. [Model interpretability](https://christophm.github.io/interpretable-mlbook/interpretability.html)\n", - "\n", "## Starting your Machine Learning Project\n", "\n", "1. Identify problem type: classification, generation, regression\n", @@ -249,7 +237,7 @@ "\n", "1. [STK2100 Machine learning and statistical methods for prediction and classification](http://www.uio.no/studier/emner/matnat/math/STK2100/index-eng.html). \n", "\n", - "2. [IN3050 Introduction to Artificial Intelligence and Machine Learning](https://www.uio.no/studier/emner/matnat/ifi/IN3050/index-eng.html). Introductory course in machine learning and AI with an algorithmic approach. \n", + "2. [IN3050/IN4050 Introduction to Artificial Intelligence and Machine Learning](https://www.uio.no/studier/emner/matnat/ifi/IN3050/index-eng.html). Introductory course in machine learning and AI with an algorithmic approach. \n", "\n", "3. [STK-INF3000/4000 Selected Topics in Data Science](http://www.uio.no/studier/emner/matnat/math/STK-INF3000/index-eng.html). The course provides insight into selected contemporary relevant topics within Data Science. \n", "\n", @@ -257,13 +245,11 @@ "\n", "5. [STK-IN4300 – Statistical learning methods in Data Science](https://www.uio.no/studier/emner/matnat/math/STK-IN4300/index-eng.html). An advanced introduction to statistical and machine learning. For students with a good mathematics and statistics background.\n", "\n", - "6. [INF4490 Biologically Inspired Computing](http://www.uio.no/studier/emner/matnat/ifi/INF4490/). An introduction to self-adapting methods also called artificial intelligence or machine learning. \n", + "6. [IN-STK5000 Adaptive Methods for Data-Based Decision Making](https://www.uio.no/studier/emner/matnat/ifi/IN-STK5000/index-eng.html). Methods for adaptive collection and processing of data based on machine learning techniques. \n", "\n", - "7. [IN-STK5000 Adaptive Methods for Data-Based Decision Making](https://www.uio.no/studier/emner/matnat/ifi/IN-STK5000/index-eng.html). Methods for adaptive collection and processing of data based on machine learning techniques. \n", + "7. [IN5400/INF5860 – Machine Learning for Image Analysis](https://www.uio.no/studier/emner/matnat/ifi/IN5400/). An introduction to deep learning with particular emphasis on applications within Image analysis, but useful for other application areas too.\n", "\n", - "8. [IN5400/INF5860 – Machine Learning for Image Analysis](https://www.uio.no/studier/emner/matnat/ifi/IN5400/). An introduction to deep learning with particular emphasis on applications within Image analysis, but useful for other application areas too.\n", - "\n", - "9. [TEK5040 – Dyp læring for autonome systemer](https://www.uio.no/studier/emner/matnat/its/TEK5040/). The course addresses advanced algorithms and architectures for deep learning with neural networks. The course provides an introduction to how deep-learning techniques can be used in the construction of key parts of advanced autonomous systems that exist in physical environments and cyber environments.\n", + "8. [TEK5040 – Dyp læring for autonome systemer](https://www.uio.no/studier/emner/matnat/its/TEK5040/). The course addresses advanced algorithms and architectures for deep learning with neural networks. The course provides an introduction to how deep-learning techniques can be used in the construction of key parts of advanced autonomous systems that exist in physical environments and cyber environments.\n", "\n", "## Additional courses of interest\n", "\n", @@ -271,6 +257,231 @@ "\n", "2. [STK4021 Applied Bayesian Analysis and Numerical Methods](https://www.uio.no/studier/emner/matnat/math/STK4021/index-eng.html)\n", "\n", + "## Hot Topics Now\n", + "\n", + "1. Boosting techniques and complex neural networks\n", + "\n", + "2. [Adversarial examples](https://medium.com/@ml.at.berkeley/trickingneural-networks-create-your-own-adversarial-examples-a61eb7620fd8)\n", + "\n", + "3. [Zero shot learning](https://arxiv.org/pdf/1707.00600)\n", + "\n", + "4. Transfer learning\n", + "\n", + "5. [Model interpretability](https://christophm.github.io/interpretable-mlbook/interpretability.html)\n", + "\n", + "Based on multi-layer nonlinear neural networks, deep learning can\n", + "learn directly from raw data, automatically extract and abstract\n", + "features from layer to layer, and then achieve the goal of regression,\n", + "classification, or ranking. Deep learning has made breakthroughs in\n", + "computer vision, speech processing and natural language, and reached\n", + "or even surpassed human level. The success of deep learning is mainly\n", + "due to the three factors: big data, big model, and big computing.\n", + "\n", + "In the past few decades, many different architectures of deep neural\n", + "networks have been proposed, such as\n", + "1. Convolutional neural networks, which are mostly used in image and video data processing, and have also been applied to sequential data such as text processing;\n", + "\n", + "2. Recurrent neural networks, which can process sequential data of variable length and have been widely used in natural language understanding and speech processing;\n", + "\n", + "3. Encoder-decoder framework, which is mostly used for image or sequence generation, such as machine translation, text summarization, and image captioning.\n", + "\n", + "## Reinforcement Learning\n", + "\n", + "Reinforcement learning is a sub-area of machine learning. It studies\n", + "how agents take actions based on trial and error, so as to maximize\n", + "some notion of cumulative reward in a dynamic system or\n", + "environment. Due to its generality, the problem has also been studied\n", + "in many other disciplines, such as game theory, control theory,\n", + "operations research, information theory, multi-agent systems, swarm\n", + "intelligence, statistics, and genetic algorithms.\n", + "\n", + "In March 2016, AlphaGo, a computer program that plays the board game\n", + "Go, beat Lee Sedol in a five-game match. This was the first time a\n", + "computer Go program had beaten a 9-dan (highest rank) professional\n", + "without handicaps. AlphaGo is based on deep convolutional neural\n", + "networks and reinforcement learning. AlphaGo’s victory was a major\n", + "milestone in artificial intelligence and it has also made\n", + "reinforcement learning a hot research area in the field of machine\n", + "learning.\n", + "\n", + "## Transfer learning\n", + "\n", + "The goal of transfer learning is to transfer the model or knowledge\n", + "obtained from a source task to the target task, in order to resolve\n", + "the issues of insufficient training data in the target task. The\n", + "rationality of doing so lies in that usually the source and target\n", + "tasks have inter-correlations, and therefore either the features,\n", + "samples, or models in the source task might provide useful information\n", + "for us to better solve the target task. Transfer learning is a hot\n", + "research topic in recent years, with many problems still waiting to be\n", + "solved in this space.\n", + "\n", + "\n", + "## Adversarial learning\n", + "\n", + "The conventional deep generative model has a potential problem: the\n", + "model tends to generate extreme instances to maximize the\n", + "probabilistic likelihood, which will hurt its performance. Adversarial\n", + "learning utilizes the adversarial behaviors (e.g., generating\n", + "adversarial instances or training an adversarial model) to enhance the\n", + "robustness of the model and improve the quality of the generated\n", + "data. In recent years, one of the most promising unsupervised learning\n", + "technologies, generative adversarial networks (GAN), has already been\n", + "successfully applied to image, speech, and text.\n", + "\n", + "## Dual learning\n", + "\n", + "Dual learning is a new learning paradigm, the basic idea of which is\n", + "to use the primal-dual structure between machine learning tasks to\n", + "obtain effective feedback/regularization, and guide and strengthen the\n", + "learning process, thus reducing the requirement of large-scale labeled\n", + "data for deep learning. The idea of dual learning has been applied to\n", + "many problems in machine learning, including machine translation,\n", + "image style conversion, question answering and generation, image\n", + "classification and generation, text classification and generation,\n", + "image-to-text, and text-to-image.\n", + "\n", + "## Distributed machine learning\n", + "\n", + "Distributed computation will speed up machine learning algorithms,\n", + "significantly improve their efficiency, and thus enlarge their\n", + "application. When distributed meets machine learning, more than just\n", + "implementing the machine learning algorithms in parallel is required.\n", + "\n", + "\n", + "## Meta learning\n", + "\n", + "Meta learning is an emerging research direction in machine\n", + "learning. Roughly speaking, meta learning concerns learning how to\n", + "learn, and focuses on the understanding and adaptation of the learning\n", + "itself, instead of just completing a specific learning task. That is,\n", + "a meta learner needs to be able to evaluate its own learning methods\n", + "and adjust its own learning methods according to specific learning\n", + "tasks.\n", + "\n", + "## The Challenges Facing Machine Learning\n", + "\n", + "While there has been much progress in machine learning, there are also challenges.\n", + "\n", + "For example, the mainstream machine learning technologies are\n", + "black-box approaches, making us concerned about their potential\n", + "risks. To tackle this challenge, we may want to make machine learning\n", + "more explainable and controllable. As another example, the\n", + "computational complexity of machine learning algorithms is usually\n", + "very high and we may want to invent lightweight algorithms or\n", + "implementations. Furthermore, in many domains such as physics,\n", + "chemistry, biology, and social sciences, people usually seek elegantly\n", + "simple equations (e.g., the Schrödinger equation) to uncover the\n", + "underlying laws behind various phenomena. In the field of machine\n", + "learning, can we reveal simple laws instead of designing more complex\n", + "models for data fitting? Although there are many challenges, we are\n", + "still very optimistic about the future of machine learning. As we look\n", + "forward to the future, here are what we think the research hotspots in\n", + "the next ten years will be.\n", + "\n", + "\n", + "## Explainable machine learning\n", + "\n", + "Machine learning, especially deep learning, evolves rapidly. The\n", + "ability gap between machine and human on many complex cognitive tasks\n", + "becomes narrower and narrower. However, we are still in the very early\n", + "stage in terms of explaining why those effective models work and how\n", + "they work.\n", + "\n", + "What is missing: the gap between correlation and causation Most\n", + "machine learning techniques, especially the statistical ones, depend\n", + "highly on data correlation to make predictions and analyses. In\n", + "contrast, rational humans tend to reply on clear and trustworthy\n", + "causality relations obtained via logical reasoning on real and clear\n", + "facts. It is one of the core goals of explainable machine learning to\n", + "transition from solving problems by data correlation to solving\n", + "problems by logical reasoning.\n", + "\n", + "## Quantum machine learning\n", + "\n", + "Quantum machine learning is an emerging interdisciplinary research\n", + "area at the intersection of quantum computing and machine learning.\n", + "\n", + "Quantum computers use effects such as quantum coherence and quantum\n", + "entanglement to process information, which is fundamentally different\n", + "from classical computers. Quantum algorithms have surpassed the best\n", + "classical algorithms in several problems (e.g., searching for an\n", + "unsorted database, inverting a sparse matrix), which we call quantum\n", + "acceleration.\n", + "\n", + "When quantum computing meets machine learning, it can be a mutually\n", + "beneficial and reinforcing process, as it allows us to take advantage\n", + "of quantum computing to improve the performance of classical machine\n", + "learning algorithms. In addition, we can also use the machine learning\n", + "algorithms (on classic computers) to analyze and improve quantum\n", + "computing systems.\n", + "\n", + "\n", + "## Quantum machine learning algorithms based on linear algebra\n", + "\n", + "Many quantum machine learning algorithms are based on variants of\n", + "quantum algorithms for solving linear equations, which can efficiently\n", + "solve N-variable linear equations with complexity of O(log2 N) under\n", + "certain conditions. The quantum matrix inversion algorithm can\n", + "accelerate many machine learning methods, such as least square linear\n", + "regression, least square version of support vector machine, Gaussian\n", + "process, and more. The training of these algorithms can be simplified\n", + "to solve linear equations. The key bottleneck of this type of quantum\n", + "machine learning algorithms is data input—that is, how to initialize\n", + "the quantum system with the entire data set. Although efficient\n", + "data-input algorithms exist for certain situations, how to efficiently\n", + "input data into a quantum system is as yet unknown for most cases.\n", + "\n", + "## Quantum reinforcement learning\n", + "\n", + "In quantum reinforcement learning, a quantum agent interacts with the\n", + "classical environment to obtain rewards from the environment, so as to\n", + "adjust and improve its behavioral strategies. In some cases, it\n", + "achieves quantum acceleration by the quantum processing capabilities\n", + "of the agent or the possibility of exploring the environment through\n", + "quantum superposition. Such algorithms have been proposed in\n", + "superconducting circuits and systems of trapped ions.\n", + "\n", + "## Quantum deep learning\n", + "\n", + "Dedicated quantum information processors, such as quantum annealers\n", + "and programmable photonic circuits, are well suited for building deep\n", + "quantum networks. The simplest deep quantum network is the Boltzmann\n", + "machine. The classical Boltzmann machine consists of bits with tunable\n", + "interactions and is trained by adjusting the interaction of these bits\n", + "so that the distribution of its expression conforms to the statistics\n", + "of the data. To quantize the Boltzmann machine, the neural network can\n", + "simply be represented as a set of interacting quantum spins that\n", + "correspond to an adjustable Ising model. Then, by initializing the\n", + "input neurons in the Boltzmann machine to a fixed state and allowing\n", + "the system to heat up, we can read out the output qubits to get the\n", + "result.\n", + "\n", + "\n", + "## Social machine learning\n", + "\n", + "Machine learning aims to imitate how humans\n", + "learn. While we have developed successful machine learning algorithms,\n", + "until now we have ignored one important fact: humans are social. Each\n", + "of us is one part of the total society and it is difficult for us to\n", + "live, learn, and improve ourselves, alone and isolated. Therefore, we\n", + "should design machines with social properties. Can we let machines\n", + "evolve by imitating human society so as to achieve more effective,\n", + "intelligent, interpretable “social machine learning”?\n", + "\n", + "And much more.\n", + "\n", + "## The last words?\n", + "\n", + "Early computer scientist Alan Kay said, **The best way to predict the\n", + "future is to create it**. Therefore, all machine learning\n", + "practitioners, whether scholars or engineers, professors or students,\n", + "need to work together to advance these important research\n", + "topics. Together, we will not just predict the future, but create it.\n", + "\n", + "\n", + "\n", + "\n", "## Best wishes to you all and thanks so much for your heroic efforts this semester\n", "\n", "\n", diff --git a/doc/pub/summary/pdf/summary-minted.pdf b/doc/pub/summary/pdf/summary-minted.pdf index 54bff995319b11734ddb795b964b23eb4982ef8d..50f6829b4fa7c1cbd5622ed3c647a0bbcec6bb8a 100644 GIT binary patch delta 115866 zcmZs?Q8wiIwr$(CZChvUyY|D4IOlErfDzGJ|N7UP zw!1+i$9g~mfS^DyAUF^L2nmD&LIYudus}E}}`lDY6!S{E5fNM>Gg)2stoH z$H?POp0jQUhCGu0-F!ax=X_>wcXN4RzbWitTBq`~z*C32GQx{c%d;>>qn$Ln2y_^J zl4N{2e;E5*AL_?m6U~tQaiqM*1iZ6+^kbDRpn|-Qrs%n zMs>bf=4vgMH_KC@+pr7IUbHbi{bAZ}AQWN`4bmwU@G!G;u%_*!9!h*UVuDxc*Ltt< z)*so>sxxwo- z+(ziL@&58?Jp*!Sq>78p+EM@)BkPHZ4SaumoRN7^0u6kiC*-)6tNulo%A#XRL4*F; z?aS#@n~!8OcA^=&7x+eK|qJuSH9W-N`?D&sAesboE(P+>G5NEVYyhjSl$6ysf+0u8Efqa*{Z zwY7vn1058sY70~ohQnM$XVm%kJ}H=Mh|QQ@qCy}) z3y z{%d$j%s1g2m_398#)RYR7)%J`AbqYJU{&<$}0-0BZw(?xqKlo-|9WDsJ!_^_uo6yTUBYZ`z{J; zq4o8k@ZJ$@Q{D@}c55jgtPp_^3wa(cntmvGs4m{um&)$(^ib!iDq{Qh7KmtzUFu>L zP^)hB0C1T|8Q#5SBSP)JtS9(zAB5+otoEM-LZ+89R-PTX}3uG%VmcyOD@{e3@ z?INDaA5{Cl^|nnc?z`;h#um3QZJ=z07AUh({Xg3A>8{=YtapE7O*9zmQ}HF_f~USk z-`B8b*&*l&d+cA~%YE?|@#R}6H2X7!TWWgFpwnz-QD36`ETG?Bb>zv@t(_bJkz_mk zO=~wrM^;j_1~w;@%8@Zpy=}hcH+)45`F#|&5{`QJ?CcR;^920kvY-6tKZ4zFR32aKCrF*OpXeRAFvI$OBH7- zOfn5|M5eSIn4OztSuj3r4vD_}d#Y?Pe0^JG*tP=R*~>HU4xm~3(l#RqDFkJ+zHI~r z1A{kKyve*gWN{tE)Wf)T&Qg?schqA^VDOhiqCGU=dE%iW1Ke2@>d)qB(aI07dK6?J zk!|^eu(WQgxqP)NU%2Q{#o_EB6zV`0t_IspYmXjr$4#qA*(aB<9RJLb6ms!PH*s3^60pE5TcU|MTI) zGGi)W0GYeU?`yd%W+CrmcNd(R%y-P}BSN+_Vq`{4Z)H5ppS~Sc+H`?+yg_vQ*I+V4 zqn~7(xqyO+@-C&9SyhL!Tq%8?9a~awXqGeMX1yYhF)_GZXASl34I#)w0s5|&xKtT- z2EFz!coStmOQD=Ao(t|s#8p{SF%y3&iBBu zjDB`U#+fhP>hrHGdd$JkpsDQxnkd=gfac1l**i62P?n$%&UWg2|6BkiG$Nn*^MsCl zicHH&1q>*BRk)@{v*%4!Z5|DBhdfFG1G@ZF(fSR0eK)?SVt%^LK{mf~s<(8|8_)Rd zU(qh!ZJ(2?tEe8((e@GoICBRx*FP@i#`gd7`XAJBa{u4V<7DDU*CToS438F=xiHeLOO0e{?MZ3Ts8}ps!h~LBez*Mdas%J zkTP>41#FeX7Lp0B>0~1DJTt%2w@CZ7RIo5{w zGxjAK`?ff@xg4-%WIHYa`8f@WihEQig+gvB*o`{7^5-}ymiEBv13vk? zBX7*1#(t3I!QoqQCs&P7V@N(n{Ix0_>qw`0E?`<&@bE=u{KEBs?#l}tGsLK zN8-|)$hLW%#!Wa^tsYZ=AeEPjI!_iATO$>EUJmt zkuv^lfRWCWxnP|lnz=?*<3cbo|D!okV03W1nR-M_7Jm!iUVxV7SfhOV0~&Lt9L|ti z9mS>Y4|LWpcS*s~@|<8Fw^^${>GC>k@`FAMKCd^8LPF(u0M;S08_p>omsSxVO|!9> zW|UMg!qGCPN6}9o>H<*^rtLJ*j&Ha)t)bLJDgFpPV8}ojJ5RY25$=08dqYWuTGygu z0ik5U!|y^|luE?b3m$zb6p(EP@pCcq)BMPWkbM3rjM55LT@~WBlgZf8YtQ>BVW|J= zY|t1=o!B10CH0;=z`|r14FAe-Ww?{1vATP(nV~SkiUk9Eb^UVH zB1?cFOz`u2s3w{prm`M9?ocuTJz8?K~`N{AmjWf*>^8wtKxF>Fo`4)svMG=24 zVir5}H>?_`{?sV$PaY^|rjbO%mt-$I5^y|64uH43E-X!rG_PgxlJ_H`CQ+dO?1jT$ zjLLM-Kqd|%5P|R${x)uF39PGO4n zdK)H!_!ax_?G!Kr0n!I6!2j5_Ljn%Ma5U~9B~{*QP6n0oiy^)K^ArIaSNZdBNZw~U z5r{fAG$EIVuDjZ^8ZXjbk-!N+7M6{#wO!^9W!yQfuIL`t6Po;7NLqU`=Y&p7LLQ?r zrhV$JGLAAb-oZ0n5^;(_n6!7hs_pB58r_N)r@%n2`3tzSwS!uhKi+UyXoR9>M*{IO zzIC?r{z+z6NKBSTa>GoAL>vkFQ`^t&WaF2KI_<}njWS`uHl9OiE6In1j{BFiNHrgd z|HF)@Ys#yT` zwc`_8K^5T4G$K{AfOx(5B0nrdPNb_$11H_JX0M^<4YTOtySsXMo6`)6>oPTufYoT7 z*_2^^PDzk!P}dvwYga=u{+`M4+1}FH(vi;J{Zsh`ucn=VjgQmwU^+Q5MAHa+vAMqd zE8#_=r6~k?r7eG74`K^TV6jJ&*kbmtbg{%Eng&3N#Mo%Q?xGopj$;NAd$4UlNpjfh zjDYnS4zo^%!yu^r1@!eRM`eqWhSL`G@>zT7xPEeZdwrE^W?$1pEXyb=i;`bTlFdbJ zO-hmVTeEMM!O=-_E^ zpJ!xwlSK#XD=}~-S*_O59Ce1n%8mhn#v2& zH86n3LCKXpl)6y&&OzR%ug&cBoO`zNGY@84hrpwkgxgLE;r z%t}m|OtFxu5brtXcpLHgF?4qiq7tTR%u&_*&2>}w1&_5{Sfv~|bzPERfxYP#G2!d{0^2Udv z!x36uLe*|Pj?5uuuLdha2G|0z#QbHh;4*;-S;4|uLipOSi(XjDR(T1&2bf zhE|IB!o>5JW5^VgnB*2W{_n zbPTJI)1f~Uv8vqq?;`?gw$fJu=>$Tlf@#Xo!o3i~8cm^5jM-7dWuUVt z>ncK$nlt-hKkImfZ)1a~n^};R_eTS$gWG>*J61~^GN#4bcY}Kv*L0=viLrJG4jX;EVXiC4HMF_2BL-AuZJ8?(G}L zBgB~5lwwQGH|R{e@pcN6GO0KOGZ#1ef0?tSt?!6Gg6VfsgOkg66>M=lJ((5LE5FC( z^qWE+Yyxr~E=V}R)><`AmM*3K;l}4WzMf(jO(GF-5qX{=%<^W_=lX7`0JKI%8V(SH zHb!d-E;}x%M><16;HAml#)+ADNx=Dv*#4=sTAR8TN!xu#eDSM4YS_ z*a?$OcPO4}REI-|&}1NPjGPhmr3R4NQ@$~>1D-K}LZjE^lF{61$;G0FQ^C4`HOR!G z%Yato!bngl4d8N(M&+WF?=x!zyo>e1n(wlTDuCwTy2!vJAaFza!eL@bShmVvPUgZ5 z4hXmQTZKb)^_uG$`$C0Oi+Lhp1pB6<_#!3j3$ee_#bW2$1@aZ=N}LNT1OYOwgsX={ zAn;g3LZ2Wg*Gw|V3fJxRn02!;YSRu;XZvt-jNBVhMgB3|yqd|%~Kgp@D z0}8C7UEmTdXA-%Uk8idv6aZ;AOJ$PHohrxXyqRT=Jwq$#5AQGP1(S>m4;Cf{e3m@aVcY}LR z4@M%G4xcg6?X85v^J^k!g|~Mie})78W{#fk&*v74=B;(ij`%F{zJG36TT1G1}~ z!7exrZ$g#v;@PPr1%N+Zl^o2^?COZz++O;bf31^{Wd3uN;m0;x(1*jpD0?QfSvMX1 zp+lBh8|lxngHWuWE;GvQoE_L$rerK<!8yrS?hIZK zAIz8Os}VEKa&Ml5D-%DPE_Q%l43oo+n9`1v)OMc+=#}@A9h>IU(s}1jpZch;h98!x zrid;xgz?acPi?Ckiz4wH@hphYgkSP#lBlq&k3T_b$6nA3(B;hqU)0L+B=KPTI308C zr0BQs_Fl;i08U4r_XxbXJYJFMQ{HOdr#Ie-T8e^O(x!A(x%tYTO#>FOz-t+ahKR^* zmyHR$x~7)}tDkF>KTE#Co@G3~+O(&3aOF5rGS+N-+tv4w?UV#Wwa*SI&o5i%l0?4& zb7<>aX>nCcl>h88uF&_ODY^L!#+LppJ=%8JY1;B<0QkijOJG^6CybOWEHwW}_&7Ov zO+weMRHd8m=rv9GblE|ZeqoL8L8xwkt>QS$l zN43^0tcaJSxezWFP&=0mpt;0f9IC*=-14*~&R&cz8RL@6ltvNuOC`OywVvg1cHMvb zd%Mc;*5(w~?bkuHi5SME?EPczPIX#~{9*0_w?c?iv{5ZaI;6S6zw%KA-!E^g6nYK=8&eR*o3 zN<N4y8n_V3=UN6qPOUd+7+x_-6ioOGJ@De$E zfN3~w1J>(dy_kB7vjAqz(9j_j$8`lp-UYfD#*L?tvkW)ZD@$yP_vQ@*3n-%Z<5PXk zjY@fV{hH!}*0(>YL{*km2$B~3cE|P&96823$*%u8mzxzNXA0_)A2b~{pBlsVmJTvq zuDEdx#*%`IyRwXRkGvP?l3#mLktWVa04eaU_wUeQv%%UA#{;M^s{m2AQdC%p*grqH zxKk5*e4@X6%_T;y!Vd-|6~(xB2yiTrwdcBRr|uND@3Cw9ET>9;L4+mB708E|)=WU* zGuKiOM4CIn!q7&$PW#mv@$h`ZP94ERV6X^HD9Rz3wneAchJ9f_h}U{q_MfN-0=VdV zjv-C54omol>(aMQR(_Jng8-w*I5daUwd<%J$RD@SAJ(`0qSqyT+vodbXm-kGwCYG9 z?8ALBOsr6yDzMMdu#9R!(*ELQi)$r?e)Ln-xiUpMPJ^Y|QEmpTB|NpA$B||80w%bf z4uL<=`ZWjj|Nr|oN$37E0v|ox&wGfMhf4N7Ivt2kcTCXl)9`rBt+|4zz49AFk{o9#%J)Yr>VW%(o zy*#F7+a%{l$zLUHy)wH5OL4jdy9Pl!|EPP~z;@G{0nfei_+& zwqs>I4SYs^tF`#%bd~q3?(84DaG-VPRvFtvi-R~2Tho&{b3L3(I%Foz!TJR}w$Pi}T)S`z?sq^0;b2ZKhYqq*jLl>DBbQVSq`SEREZlUPj^d z_WYGmIf;&;BSDOohWdo`R7T}z_UJl0Vmm^WJh#yxa37(W+JJSvW-B;ouD%p?V~PJnd?i1SH7adn9Rlx$5hmBBuw|VYbt8vHdMe9e_*{~ z#d$iK=V+4ncIeLA)1SZ0eyOL8Fdwp5tb78uNs)*-5E?w4Tv$b$S_YdQ7}L_}FrJ2A zwvs#)YCh{B!y2#o8&a-gs~oNcNDOLR{egL_LW^e+6UO>->Y7#gAXRTOf3a*3@g{h$ z7Q|0)tH7VcW$Q!-Ef0UD7R_1tvLXZ&AVb5$TI!`fF5ED6?DMg#<`<~rH|Az0oeG^xH@6?>N{vkqhPA|ylXA0YY)5Wb+1?W}`#G2RsEHX4j+#6n8)(V0TZO-7oqD|by3>0XP z3T>;n+2tjXYW+N@ykX*2`#hkV*PiA+`kYpGKE}iPCpiDJcymrNN9(v(C~pn2XaXN7 zS5cYnUN;ftW;<13ltfkOpBuz4xfKSXh6Y#WrE{^_`Y@-1Ga@4aJdQ`EhS$&{^jzgu zNzQJm$HUf-zI?s0AbF!9CHoJR1xKE1c({a8Poz`yBsG~W*VIy}5ERbYc0ooN`hB)a zxF%($BP>n+L+&BYyPP!8H;OCT)A)uhe)vYE`04owyzJT0P-Mp zvt_t9(v(?wRA}D@lxhVV>O7=@)eihAsTW%av;j>iqU-_e5(Rk!b1N5`VF}`XEdQeU z>B|-F^gaTHp9>p!pE74!oX#gET}+Cv@qBye7e>#g`HjmC<+^=tbY3sGI zSbU#4i`f|Y4NpydO@zEH`6R?%?{ z;*&o7tbEw77N&gBN;LruqSJPe@lJ4uwfBa`x@ONVgR7)D>*q$r)#hbo!Z>>WR0^dKJ6N(b9k}s?l-~#$<;K@)(rPlDa zZv#NKMV)T|8FZ;KB=Dq7bM=lVA8eKLVIv<7u`co-zq#~~esyFeH`!=l{JRjmTZQpE zsWhgrkO7sJx^d!u#whs-Mz-uWdu_L`+Gg(&i3bu$k5NYsiUO_BrX}KOuWC#Hvmg@n zM$8DhqY3Xw-LpCIRY-#(B78+Xiwm}mzgAY{sTehY0YU!!R;ORwk(Pqz5kYHD*2MzN z32LKdY0DX+@=J(@`B9&e)F_%Ndzewmkz`Y70DCp`s~rCd^3wYC_s-!+!?o4=y>V+Z zWE7SPDp3KWty%z5ev8vwNq-83CTuTa^KIs@atNbCE53{30VaueCOk4GkS|ZNN6AKZ zIRRmS(#!G}{NRMz*dLE zNy#yTPoU(6-HYw=eT;IP4@vpQZ%5wc>56Nhir@*-XVt6S?i#uA)7NGch)_sBaK}m# znGE`Z*EE2$ec@GV%T2*+k$HB*-`TkulKz9RFDN3$@nbH62{IJFM~t~D(V^L}1=$Ul z(^JF{r0;}wGVS*Z%Hd>sw?x||InROMGx9pV@l*5qxq@Xv!(WKS6^rf+FR z$%F%_kDbl@_B7cOZJmp$O566o0L&VQ(H;wi=`0Y2Ex-j%SSb3C?oVis7j@*SdrmV!n=SJ6tx z9TGK^M*PC~4U#SAMfcxen4SIq4u)CSI9UHXr2o~Hbv#OL<)PGv{C5p zVWXNsfVNZ5-Ah41dMpYb@azfjYo9|*pO+dlpvwhi!~6P7pLQS7Z~l1y)$#r{u~0H2 zc(hb4zfoT5t{A4zFwvNFR84!)h;npYw)6EH^=zOg=es?7{m44`glW?Kje#+vv78fH z0>E{9usDVMk+0xw|I&q&07ZrDWkUR%Zu`=WCM_*ZnUEckeC%uoi{A|o68S!P<-T7uDs+(lTw3>Nim#5s;Y^7QOH5oC zLwRa2#jApf(+=7DYf2L9ptSPMucMqQ5uj&A=Y(>lLr1EjT#Lu+HoRcL0|oB;T^3N~ zwVrGq*u0L-V&CB1g3VHe*SsXym?p%WMfCJYat|?7HgSkuMaM!WfK4?s-)xWx#Gdr$N!K$=HhT@LMLfO4N-fvvI7!rM+a2Z?3&X z0#OXExIeJjy*cJ|(tO|Ic!2IxuBYJLLgImUXa86k-7o!4p!-0mXA2C8FJB4Xd-`_- z_RhS4U~#1i=^7FiIFv&R|F2;=d_a<@XP2T45C#tY<+_td`y<(3+O=kQTD8$^)+g?> zwwC`^Jg~{~NYx5q_~7{^OoA$=^I=eO+TEXMmlmmU{~c1LIN#VcQ-nr3trXNPqTnQ! zx^cKgsxSG*nw}5EUy|K;w`(C8x4OA;7Jnt}0p$8niEd5lPzA2f_O-o01dy-`Gd+#9 ztxVyYn*r|Vc(QkmE*2-#C(&v|0*`V+5RXrd!o;xoixqa+Wp^`?+0z!LAnrMYe}@_0 z`Gdu?fm*VLdDV<+HTz#nTP$}HA)~8QV#}zCwkr~|nH4|6aX2h`lQY_BEPee5(N@|nQXlNGs9P2b$@hTss4lQD4SMw>XkEC9kA1Vj9qXMa69 z@x`lwBN=tKxjOtw_%n~>DAYWF$WU~7)2b$*^djP2F1y?}oI^`l%<-iw-Rjj=!-CqC z&pN(DDtE98gxTxK@cCh<4A4(kMIl{=9sE|s>3SGp#~oWv_6!fj6MUuOp>rF#Ne(o48*NHg;; z;jmMYU;CJmMkp!4skmXWLxjZp6Si;60!5p}hLt#p{^E!#Pk_vkEAb>pNA)+JifTa+Q;-y`Nt1=8^LXd`Nw5yRc?clNh(h5`vkRiX z=e)K^%`i&02mE0E9F;0LiNm{}_&#G0DYc9_BXN|Kr61JnY&zvtq@ya!)9iUWMt|e& z(UTvIrgL`60RcWxTf|-$pO=`YWj!aAB;cR*b;$0+uaYAr#GH&Atw_;bE{uF4XCyRI z*|lad>i|3W)g~%iFzaSN&wg(fcL6lVumCM0B{RE{=CzPxieu>|J2rIneo z^2V{%$6S8&uUa|pBh0^}6ZxN9;rC&6KCpZ1wWxs-Zh&|2c?dEuQ^+jw>`7nQB&FDZ z6DF*!?c&$SA$psD=R<`m)FQbImFiVcR;A{2kz;!~*)>|Ozl23q2tn@hUcT!}<3Y>i z8C64c68us^&?yUrZ-WkL12*liF=GKU1yY5047pSY?V3N*Fu4lZGW5)f$l@tF(?;E8w?Joc zi*gY4G%BtVrrl_cwo!tzi2godAcFN81nB6aREk}qbQ0{(Wgm_OVR8!a;NF)-2Cr{- zOGRZmR&#=7)x`ocu2G0NJJ?DI=b4gme(RzUlhT+~pIUYcY-I*lfPKc?;7K+v|IP;Q zj7bas`@#$*iNsoUc2gG*N`me7TAP+l4poyqOSH@4>KQ7%X2E^3V=$B zjUSFpHKoue{IFlsSiHeq(Je9mr!^?*#!zYwfp*g99t_uv*#u!o>k@YyPk%>b7MS@m zS{g4v=hrJ+E#USVDMju%Y_$nzjonK=2W^N*rMLwUGMt3%#AN0o-LfV+F3_$p`UAb% z@s{Zdjvxe`d5ow_mfE?^BcHHy3eby<+D=l;cl)4+EOs~( zi8&xxoDO^`eelXfrpv^S#p#w5*rwz5S||=Na^uxVEQWA;I}-xd1yNxn0E~@RY>$Bq zAtxMfHwOYD16-1gE+gMMWVCfr8!mSXTMKvcH7(D zN&J=q*igm%A4w2BP}r6O`a`Zo+A3~%?X!cy2PZvd+X2P8h62ziu6m(SGn;OYVKNjW zAxwHr^C;F@{Dh#%hti%BOqd%XHnvik#_fG6qHu1Lxc~NEm_43uV3KhJ8Kk8C?ADR_ zZ(B1of4@GS_0WlCm;cx4IN1IlPRGi|^*>Jc3!p8VaL|m}{huCZTshtxMeoVj5ydti zt&gMGs&d@V@7o zX+K^&sn)#KW4F(w!h~tRFJI}FYB@wD`=E68r$NyE8=XrxMty;>xr0I>c+bjf_p-+H zKh?A`cikjtaMlm( p0*5IwT@x~3VTJ#&Qj^Lu^Cia@@DI?G8B9H$1=gER)yp=x7 zfh&fXeIw+ZubXdTk0$wfCCo}YRb7-ab&2v_NB$r*wsZ;>H@0l(e9X*XDOZ8uY5QOq zGXBR;AC^C}@T!(8f-TP!Wa$n%s(b0=FaRr_j_>n@LVvaGbL1YKh^CBco00UU#Yjr5 zzPKS!Q>R=>g4bmP`5k`d2JW=r$&bKByNKxEa_Xv{Jd}%1w~3m9Cz`-t3tlMip|VvO zf)4HJiU0cRTbHzl@}{+qg0Ctw7M0Ce$_n5J)vE5g-|)@~Sw6WB^NsL)S9w-n_5jE? z6a(9|I7y_yI2=sOb;K{Ep#)H|_~lRO%mQ&zxR1RCFlA@0;x14+9s}FJ;My}y5<{%y2WWo2V<$Wd&+P`d|4yHaskmT#r9r@Ggk00 z`r16%ZytnfbCt-gIQ;m+u`OZS>;Q16^RWV96G9^hwlD>`?L1&!mz**_V3z0aK&w(h zASr2eU}&5ni5eSH7>XJRTLs>RdWwBs5cP3AycMMbKEG-+`Rgt_JZWs2{Xic`ZtN-F z>(Quo2J}YLfsE~-^LvUF6Z?R>unNsSLR`T!~&I62{`Bj6GIVfXeb zIv6Z*3?k7x`A7J5t5u$KUX8&bYg3F1oLQC&yvOoDZeu+eyfsc+`0(AwZV z(m~PTQ5G(Im&N!?Eb^ZaT&iQJrcta)IfC2>8_p;pgs-g9(tzdA^;?s6_MD zu3;mBIYJNx()UtK*a5qeBVa4r%pB(|9Pe-naa*#(mp3z95C7C1fLKE0E{ zt5Vx@hoJ#uHXl!|5h;S7#>nqihu8HqdB(d&g?WWUMxg{K5Q*knc0$5~NKF8VHqR&ILGm8GAQJBnSa3~y$tY)x zEk`#wJNn-NWe*f|390H|N2|FnY7(z!R;ex(p?uD4K?3)Rx=gF1Abs$;aD*Vbd^4Pc zu9_r4(U6@LoqGP`T-qi&JFFG`HIq~@f`>vXK(9&;9#yAiK=jNVZUw<}WBAPTmJyNi zP8_7#ANrY01ggA-W45^{mB4A8b|hZRY>CHEuIx6UCAkU+Z7=YU+a!r+-Qp@A9_@*v46f`RNa$>=wg8Fl<(9K?6y@|eZH z;5+OXVI2tPfYvUE07l&ObRvcVi^Z+Hz7k z^0Jtq3Tw2+xNf!b~*-xfIUvyuq z<)vf=?=d|v?L^f^O1Tlq8R1GJ#{p@&g{r1GdAg&IT-jgjin%@GMn*{QQQhLpUlRYa zKO;V|iWXcT4B~lRt@=c)xerD}B?44}6jViI2f~acnB|9uGxPpY3|xkpIiJ!;|D#@` zoZ5&m1zh)Uwo{|*{t}#~N8MJU=$Kg~H1t*VYU`*ixP;9xuqt)CW&5D%nKMjMf>m*z zcMYy}7Jh>GOCi<}Fk1~nJFRG@)Zx%Ymzckh!`=FjWEwa8p!+cnEb{r;^Nuo#Gvgu> zt)aCbzqd`4;NSy3&~tw~#S*N)sCXk>I|@Db0M?AIE2de%@G4U>DTY1~QHUnQ3r=Pw zjyi1G>voGgy6bNP(yO1z2;mcLMG9nz1socgxEMS8)hXg}r2tD`5QS8C_<&GYH;qedeDyKsvs= z0R58~Cmfm?!KPV)zc3iB|CnN{z%euhF!=aq9;YDA z{dX^#8L(!>tDe2Am_RdR-05U<8LH`D0TW!&35{hVR6j``HgIH8<){xOEBQx|2(&y=8?@X=JXKojbf7)-7V;f}?}E1 zKfQx-rOX6x&Wi657ZJc)3(rAm0ihP&^deYWvYusDV7!$lW7AwKhM2C%$cc%hiRjkS zFj?eX>xlG0@NfG+Ekx4Eqf@~JCAF{Vnzd;Upit;&V)sbsl$x2LbyBTNkOG66W_i2M z{LCZNI4lI~=9T2{!}Cd~JpfX*F&Aw*l{n$XH%a z|DRrgak8YuYLjAsG5v3oB}-d6=Aad``=#bq8LFHOP29eXQpjQ!)FOxk@-TKfkg~z9 z*nEt9d`bV|;(^1aAxqfhSJcMFYw}H9{FZadrT=^4=uM7+z}Ky*&gFS4gIb-B*G7?Z zyjpGY<-C4*MXTIW_2`6_`P=nT=ZEEEOaku9jCL{E$sOn4>1m%80O3z9U|C1}VC?>W zPS5kp?WeMQ^qTZxqk;2Dx{E3;-#>QCEC|DzpVp{q;Q*3ZQe zV+Td`1&KrQkh$^|T-HdriM6h6C=zgBz?@A-V6)8+@5{UCJL06Or#AX8tNjOQhISmw zJ-O7rE z_WbF?VK^4b*N*VKix&&U>78v#J|AW+*!je~Er;fZC4(3X;Wg6t8uv#kKM*I*zjhBe zvC;qwA&E~Dgkthoz6w`8&-GjTy3b$~q&E=}MDC8tQ-%Q-P>QHY+?gNwZcH&u>;l4{ z>t>o76r4~Ka2N7^VbzO{#OLi1I@A26>=31ZyOV$nG1n4L?N+JE%srN2<^(8jqCZMVku&Ueou4pEsGJS+AyjD9IAl}k*gNa(Zh{>1R{H6O$?`r4Fa-?w8t z@5aK(-NdB~$mhg6a>Km{!X)b%W9ht}?m$tY^~Qf^t$FSCOF}EUkg?-vjMJ{XpN_u>S0EH7mGwyUELgkb*-FY#{TY?V307Op2LVFA$Jf>GvO>vR8$V%%5qchdd01(P;dM2nQR z+(d`G2gFO$B)ACSK>~z%&7kJ-J*V90n2v~jj^u_50tP1NT_Z=!V{!aB) zUnV$wIVYE&Q{g=cyl-g!W8h%Bf@xN`amKO`0JFGuPG`m<1VLS6Y2Cv)*gEEGcNiV2 zWSN^1CpdbisC}0TK*%73_3#_%SJd9f?=puvD)fekmtcen+0ojzRrfLNz@%ROvy}~x z#95?LaD~Y=a7n{>zPdYRf01qR@V|IjccZ+$pSQ{s7*1TyS9}Kb#;dV0A}#EU*dyYtv4jP?}Mm)9t@-z&gKb5py1Lfx*zlsS_`ZH z5l(*`DV{KjvIUR<7%_Ek(aQOX(~!M2oydHd{T75f-Ku@;qne7*zaoP$UTwBSIHb}b z#+~sGP#Eq9**sCCH4Uvnpu7mSe`$Uf0wnt6f{Sjgiyuet)Rleto$FcS6{W@aHkLwNF%lbx&^f0+#}vHxFcJx@iSRqM1+=NtnUdj)6tf^R1&4+sVJ`QL=Z{Eirsu z=(PF9!a#-s(IA)KB(St1O>l>#t`MF><@b;4J+#dtf29Yku8kmrGgTerCw9Ej0Vp>J zkUXEr?|YHQbq@zdXQ4=l+*`EbyosFBr>Poq9Arh?(NwF34u2t>1`tU(0=fQ1kG0Rd zh_d`TP+I>_m~#o+f2Qpl)YSvZjz6r@eXaMe`|pta-2<_IqsjUYq`#FDM-Xutj`)de zK)^%=0YXJ;7E&iLG!NVf1C`p*5)cfs!WLyquNU*|#^iB3$#0ZuEoZq8E!eP6>}!b6 zyV;@4cII1lCgXoeHc7%Oj+zNiNC07ogrCV2hkp5TCjYoq=8l3K`{f`jrC~Mw`N~|N zhy$^)*8+*ONR}6sa>1s5xPR1&e6Kpk_3LeM5N2)mPv32d<*Y;kNB9CUETEF<2T~5- zOxOzDY;TVfAqp10>G@;x0R+7srSc1N05Z02e~)qh878eg7;_vke;8CZU#ONB+A4v55~++9m@JCDDt6dAHj{RGtmq%o&mOt84l zwzPuW&xCX!ib1TKlg;X%F1=DuFqkwo?d3!QiSi)3x$og)cX>th}{Eq^Bc_c@g|m5r~<55xkdYs#0Q4xLoR2%nR4U004-}p4TqvQggT^ z`w?h#k%0zQ9>l=drIs@dLn+Ez(rRtY4xfO9hS8F<9jH?5hwl-j3Utmdg+(7!8%fO$ zIh;_=kdYMp9@_d=nd>2;;z&eql?s{4aW|D9bu44=_r5eFxQP@oO!A(*Uzeykui)Uh zlJ`5Flh6IED80pQJpg_Ly`&Q7BzQXG@a)127zK=FX1ZsM&`dX{5(4Eo3VzPfRLBwf z?SD?BkU|%N%_w#&1qAwLve#NhNQ(Rk#mJDM#qH%48*0_?PPsSJtTcV7lK_7j|bv@SD9t zfx;mPBk1Hk#6JF@MMJhPO{TlFRa_~ImsE#o=zA?C*a1HmOXLC8Hu|hAR0==S8zT~H zjX~Z@b|mD#^`VQF2I2DhHfavvoC_-bF|@3&!4P{xCDq)#mYLCM@IFCeApt`V3*Zko z%~{sUq7GTS-T)APhuo6^Z8aFLcDK)?6|ql=p;9;2fxZW_o|HW_*Z@`d^b>8!bH?96 z+_?C7qqoZa4Nl;ewBZDPeRUBqGv>7sR`M+yXK#RzA!*{yya{Oq%>yONddl9G(Q zUg{6*@phIIhNqT^LFb7dPw%A5@G9v zWXGt-V_#I2K*%NcSo?Qu3=kg_g%w^J59fbBZ@rnxR2A|7|F=it=}Ah0|N0&mbBdxg zDJCe>|L#Qpa|nIGf#yF`b6bFrc`8+CCjFuEXs3w1m5(eA{%G8R962;-=jJS4ZarOd z!->t^jL(P#O6I~vfMnU-HoJE0m0oDzlrP9)A+yY3okW27%u}H`R-T62EV?`_F+cp>_x5O zw~TbJI1UYa?slmbzdXRgr>>N=LJ}WA?!P8;v`?o?{ctv2Q&D(qwz0mLf5_EuySm3ya_fLan{B_HP!Nn>g=Ra3!)ul}-3C3X|@veECe>eX6B4(QN@jBzTZBH9*{gXQIno0s)AQHxC#hz9NNwrAWIbR3$GAsaU z>Bf)Hu8DH#2AM?<0Ffm`jGg3VZ#IUhutl5#LuIyP9ww?^>yiQPYolvu^lZ{Q!K+T%cl_GEB5Y0OT!|f83ESyM1be3cRz;~Sl;9_$oPRjRn_Z?csng@7^5in7fiWOi^fl0Y_}p}6-uh)p#VWR}ArK(?|T%1Z115Mg_pOOxByu4iuseNtF?^0^N!)B?XC7lo0?y zaawKUC>bRQGm>i%q!zU@qL5l!a&aVn81LgKTEe(s({o|vLXq`MS;EgCNa4UNLs-L} z0W&?yob^UrlAlvrYFTt`%jX&k{e87r``k8#HABjoc|CFpzqSv_C%a&Sz{#beDNma_ zRN1V$e&$Y*PI!150yza&1;i{M%vMt1Q=TtR`WG^*<@VrGd`tD=P)}kS&{Y@$qM0e& zE&3c#o10HQRy@acE`PnOge@<~d*uxd*#7IF(m0*)>i&vV7*TAP zW(fqv^k#b}rQ9ALJmh!dy}R1=vVj922~$4xdJ@Np=Fw}3EcteZBe&SY5ep>c)VzG7 zq?$__Ko@fA)ToSw*M;SKFg)xxvxrU ze`N6&ErBYrkj27PkD;}A(ZFd~F$4y~L(y+RkZZid0>m^Va8Gf#_ecu|LpFJudksfM zVM+(hvbx47v;T@Bp&QKu9vNa8#O2G@4`_dpfuXPOvuOdC8PEYtLc~m;^E>bNaT9_wOh_jRM&AKou8+-fBM9f3$bY&S%L~Fx>HSKkZUFBlTQX(w1cQU6lvbR<64?&T1vftZQgP-l4upL;4hpqd71JzeZ}&Y!YG2LU zXx~wQa|PM*2X71jpvVFI+MwruICP_K>TVXe7uRMVSDexy<9tKZuc>V`J4 z+Yc_zP!%2G{~-Af(*Gd)5Ay$@_z%kep!yH$|Db7B(V-Rn{x9HvfmY2wbm)W-saTl) z|FRl05eqXr=l|=L!a~Ht&dHSONXLlvgOfFtd5D1u@G?`)9B-mWCTjtQ*=lM3Z}CQp z>|gz4;{x~aNJZY*%-Xq%+qoMNc+T#e#(uuWg||GK(<(P#u7_crJQ;=LlQ}0CR@}s==y_=!1W;zA8X_z#SoYc(lH(-VHWJ zw$nhrik`BqkD%*5%R&9;`c_x7GXi=iC!@DE21nBf4lLyz9PTj%)!<|({c%TTac+RG zGw24;FLo?U-uY)I9vzTxwVmk{5L0?Ne|PyFL8-le0Jq(?#B270y~h`4?Xxh~cd6BV zwWB{8@(ZCIT~y}RShAXG62uCQ)*3h1Wz>c27Fbqvrd}pU2(Y@<9Z-b+O#%%b*22Z} zbxQo)%`gl!zZ#tN{d-P6UHkot)e3w*$6&9va(RB4aD2DgqQzw;7ADruchmd;HBaH_ z5xmIyx=1VBG~nPS_M~=V#!%x~_1ImV9SztjHj%i#xcbm|$npA>SGNHoFYXS4hDqyx zt5;y`UPMBqv+l7Z?+6T0TE7ZVTEyi;^#i4AYlsKt$B*0UtTMbR^@1v#Za}StVtu`< zBdB!vm%>Fm%#}k1_Y8IqGzbrf{}shgctHEJFiFt7Nbu+x<>mKk&;_WuIETCtp$zNJ zJ6vXJVmvMK&c(@(hugQ}eL)mFG^n11xq(-`gKH>=tLiglcXJ|;KXVLVSwP^g8_2vThQX7+bg_%Wn~0+^#}F0%fD}YGgBL@{o5R; zZ;U+950T`%Z^ZrsjD{gJ5+$d@isAZ_6*lQhGlx66Ph z1-POE!AGNg2VS=tEsgDu{whk;RO zL&Xo=P7lSe|82nQ|MpMZ&gF#{&OI*)aiHEWmd*KV{dH^R#o?igED<237<&FH`u3Lv zf!q_`dqoUm-ZLgn1iQ{=2ox{RIYesW`ns`Aj>0LHQ8=8~^WqjOK3Q zdvAx>?i~PD`F2|FV_B^KrTN$QG0G#Rx|2J&78pC7-1dKJaOikeXZHrzWh0+%^1oY` z*9PD38Cw7w(ffPgyF^nB)5vCIOmZT=(D1|{q@l6J!EcL+zP)?GR*G*g?GHDRPQt?P zK>(<-20=bSWDC&I2T8jD*C}KRPI-$sR7e(%cbd&nH@vfv9u{UOmTTPi`#HHv0SSu^ z-tbbKi0N-i-umd1_~RmiZ8VVCRv&1hP?!H=QA6EezDvFz6hlSBnN%;^p4i6)42Z1fNiWd#&Ea`h?E&QlsI4}Y~{Ofj?P^RE4CWz60A z=+R4m1HJlq@+UwuTOe(a<+u5Z8Np41V7I*)o+p=>P;5P;*yv~pi?6mB9cokl%PREW z5rLv_WPH~{u5yGw`&O+2R($qKtDujxVoDi6Y-eGwvCOO14j72E35z$g_QhG1cYXo!EqQMR0>T3);Gd0Lfu;Hu1!e367 z_?U|OV_O%5@y81Da7}8Pm_uS;$oM-B zs>)s6X;yxx$pUvuC?+HO+AbgE{@R8S@nHw`y;u7gs{t1b4hgl={dwQ2*&L8Uk}F?n z$#)gp6?f_?LYS(+Rfr>i}R(rD(9nV<`yE zSXJrl%TMV4RF}M@4U!m)jGI?3nl2|px8fR6J{`6*j1Aw(OO%4u?#(;5G~zl|5ao~; z6w~X8At)o{#fJ72$xDu>PBxF$Z+$ZK4p#g2ykPpA#Q(&-?NeXuzQ$Ezjv_3K#ks%d zuEtJBCEfS2Xu@MrZ2%Pi(V-!Xa#D1fP#+t+NJZd1kUEK>Q#ZVXUQE?Z(A4P&i&11a zYnTx{Jei_kBd=w!?cS^YfM)m;KWiFtQe$22ld3FM&QufeM6B>Jf9*fDteynS7sOBwp{xU>q1} ztgTQy9!`~bTeklg4W7v!8AnY8cco&>u z!ah`tN)~(IBnIB0`IY55VRU^Y175TXXy}Z``9kkD5*xk1XAXtfLB4O7ejB?)GenSa zdS?LTC9#A{0(YhTHT0}7;OWOWHEbctK*qc})teB`@mnuUr_*?xK| z&{u1u@AceLFtSx8SrulQY2<}GyVQ*DPR7kzUQpbxdeESpd0}*+=wL8oWPZiyRxW3} z4#Q!--WnxTvu_n1%(avj2-YuU;5}2wJzufjm6QZ_Z2~qWYOj0G@dfUt8WMyN@Q+0Q z$tB~e*NA6xy7lzyoc{awVjwb#^lU5dsZEp=iQ%P7@6opSDGYp0bllHrU8rzk4#b7i z!Y~P_Bd>~!B^=N{^PdQxB6JzMPP%IKg#5-e({uXduFm$}NtgF;w>MQYUn(`J^1AZ< zZlcjQg-h2{;;R{Hm@%_G(aEbF93)k|AT5B0)sq|j zEgw{8D4tZ_6==KMSe9A6j*JD@#x#-664t=cFi(C|TQKSiHZBdzm>e=|8Smeegvq<` zBr(T8U`nZ7sN^mV^DUcZxRn=(wR_8D60i@_1qJxWv?rHkzoDxezcmC?Ji*+Jt-h@O zF~4(Cb38J7@w(C08(CLu>NccB|3M$3p`D&ocUynA5h2*g)j;GYJtcC(*t>MS z>~j({vSg4)wd2U~>rNCZ)wm1S5o2eb6g>4DlG7ZWGK!Ar(q(huJ7x0*-8S@aq7BHKV@{jKvvJGHhvXka9>6)?o43X%fMW z)p!#<7kcE~il2*AZTQoZtaK;ENs%)A8AwrVEZPlHUg=K1z-LPAeGBekIQGVZ9{_tI zV@5gUjD4`DAXJC!A@g~|Oc7);!k2e2_6qc$*iObUhf1$vt_UMPxMbxP5ol!+9uLVJ zM7hD)=f+%>XNL&+BF!T~4{Q@$Jwi(k@My_%9NB9vJu0r3fPYRj?eBb%YrtOeNQVdc zM@gzm4nqaKhF9*XSU-zqV*W-itHLj;J2TnT463ryZddB(<(cr+(3nn4Ry1&q?g5D0 zSZu~qWMMy>I}v@Rty?Zk+tr;_-&sd^QLbJX&!$hkCm>nz+M6i~WkTwENw^Q{_zQ4r znTz2SBc?cB-CB^{5B|*FLbH)0;r3&>oRtb)1^!l0(yVNXGc_EN8?d!+j1T0uzh{0i zl8*8|2s^}cx}d+LB{{Bmntk0>RRRW*(lN*diYotZ592RG}&!#Ii@X(Ia_9k?jvQZr{SpF}m2z zzR%2T*pVq1tZdxbBX`?g3cnIWb4K<&Kpo^H6u;SyVm(UdWOCf@RbWugP&vt!-274H z@4N1QX-7r*(li;q|0j7g^ZW1pHNs`zUs#~qmS)_6x(mS%N-n!NK%m-VTI@9&Czq%7 z4mfGTE}XXuAN_9sqo|aJd@!u~&8DQ!B&iM8mlb+_&h3OAfprvgJCZSvBEHIJ%R?gv z-kBs_p~>|3c5xj>9vsTJ&E!ty5R*e*j@HEoqvg)?g~>N^`+mIE1_PKed1hD&oN~ov zfO}iL$S=Hh&cpMRs+Y`F+|6)_tpDvHby6r;t4F?0@UXK@)$6sZsgvzvpO_(4XILX5 z_06R0xrngtqbQq=~{&0WrDF>7s6X#PJW5byi<8Em^TRnT8R$zsf(P^3-I48nOdG-m2&Rq<( zO=m==Ld*Oe@eDn4q|ghtaXgN#l`%LGFBSAd34wRUYu{qyLRe69TOe>OIsyM~{eBB-$-S(hSgU?kjh{x9z(zuoflhj46Kg-DfiXqebRS%Pa zl?(Yp!xL8?{h`X85t2o>z4=7qd^dKs*F)guwfWWfxbeu)^4>4>Eg?hZw+CqS0=pk_IgeN*Mn_=1lI$=u1y65mToekI-Xueo&K@!kj!)pdw}g(ye_E`Cgwy^IPu96h8C`#gESFnEl!6g-s>Km@2_mNP{ow zLwAh~E)5a8gOtg_DYa?6Id*J#-6!0Ycv4CJs4NgX?_6fXEuA-!40S0OT(@>1f6xCw zI(I>VwgXVbBluIVfjNFxmo&4z+4C1%kWkC+`#z1$?9dPf@9FYdYT14)Qqov6QZ1JlkhG%!(d+UJt)& z%JdwOIb)@Da)`?ZywpSMS;WU{MMdEvng2)7-$)<}s#&Z%%yNBrcl@AleZN8~9>-rr z5Ye~7mD2X}%Y@&UJ&9nbwV5U~1DKS7D)M0k*&Wc#P! zns<4IrQ8ryf+}5S%ZGqz<+5jD|2kW`W6@d812iX9x! z5fiX*Jhi9^nS{N)t@FX@|NeSJ)}3e9zB`h5abl?Q;~K=D@Ylk)Ez)Q7vOcSe^I%-t zO(67QH;4r?Fx?ZT;{UGc=jW(X_GFs0eNk};9J~iS{&R7h$mchH89$FAM{1FJr~;gO zkr}$Wrmm{}3Zk<52GfynH1zL5Ua!{o99s7^&Rf8ArJoG-Bal6i28*ut|R83$TET-U%_L65|EWA}Tfp5d?vHh0q&M-R`SvT$Pk zrd%Gm`7?wosyQc@Zfml`<;qCF#JslPBgaSda{sQfc%Qmqh+89A?0S zdp;S`*78!4Ga?;OsvTT8k8>;&C2ix?lP(N5&m)&X(rhVFnYHZO;HakLCIHa?e7?%O zg8GC{Y3#QQvOYe#{>-MR!#(*H%TMRrIuo6S^S0x=-9aV{`?94XdQ6?x2-ebmQsv85 zX`Sop*xJnqC`u|9H|*E)(FRghc6?hS9bTKnU=$L|)(Cg-ln_W>mzxq<>E~M_2{@R! z`QR)6XnGl9LSXniiN#>SC7I~+>Yj9EV;?wp9n(K? z^&3(5UF^ptf~^>DxdTNwR#V(wo7BHO_DS%iX)Ko;o>ciIjj>nVHz zMb(AscJnR5>rwHl1la%`d@Np#O*NWRk+ORlrho~-yPGdVy9yyG&FCueicc|bls+4$ z^WP6=DT#J(ZF)YMiG8HmKb4w8n=+!aJ8l!TJL9!P9Zay&mDG6a8*P)~dQ(|baC6u3 zX??JbRMxBUO~WWKm1inp%;FrY_%*0W=`&UNM24zWm!KQa`A>j1+P%%-6#Xwc$A)dP z#*%sj>sMNN)p4Q@r*20Q3q4aBKfa6o&^oHZ%nOm|=3z4Z2qY3o3C`ofvXhOAwf1Zj z`(6B?bSulyGLJx>KZGNUQA36()|}uh8R@{EwV7Z7&S+>;=q9C^N!+h4GhB2#p3dsr zRBDG^fZ`VX#S3_}!*k!}w9RR&#GS5&Fq`H~CMy}FvRz00J~(vkayJ$NMe3yjL{S|a zlqj_O1N3<(cEZ_TaH$-&8J8Z|{5;MKk-eJ#U0BkKJVWa>Se>QZ{-tNbKnh8JF;Y5M z>fRF5L)Y?tEnCDm%>TVa)=qHWGXhcN)Ys>Le8pj~W()ikxO96`EMa%vNd|?N6Bm^| z?FAzlly`MnN5uTyIJ9Hkr|+J=)XYXnuFdgXEA6DB{>UPNp*ov+fL_@k{mP#VepBI+ z4~HFdux?|Dp%jH|eY0bMS0<{C#?8s|=9^K~HS9xK2B|I#!tZt@Jw2W5ge~%>WVc$o zC1il|YZ-_&zPivLw)s0<0{%z%h}$eCye-@*j?1N>goob`q0q_qtnpV;V8^U2 z*=o1o3V3#?A*C++htQ)sWxH1~z_vrm-fKvNou>nkd zd!z$NlHEn+R$;{w5&;{GU|PC8pQW~Hk?zZ(pE5vB5J+@$+Is3@8?k{Q>cC!u*V&W_ z3T3Hx;;I0!Il!rhoHC`bT_0=y&ZP{aMciu@TYT#w)?OAyEeghI-BE!(i-RIWom^{F zP@}fUjpu}WWZ02NJ`pIOUDMrYa$<*0qT!p{*`j=XHF#bJp%9D67wM{;=P&O(YkU;0 zIs^s-xAKaO4Xev+|u&d)D4|Md}7L*X*G$%vpf z(cnWq+EaH=JhT6m)g86^Dfz+FOB-B16AW|{3nP4`J4OkJhHaccRr}9}GjP00{w5Pc zFvsK*sY&?pvcfkDF-h?U6AhAi{!~U(aUhp_BWqJMrl{cCuw-4Wzh0AcjV2D+O(vc}HEvv_WCAs#9aJmJ*;x)M8xLPWEfRAg6Oz zJ`|4Y*W2->AdTWR`#S$Y#{&etoYbiD1V<(EQT9OIBY6U{F6#D%J)=>oo1=@MbMM7C zJ425U{;kH9N=g07STSFBxkOW*HU=Ot+c+l6>66ldM@sCR$LrI@aC>7Cx)=M2ozvXS z7lKm)PCj;QIr~2pT|0V9g)4ZmzXjIZ_fQ9+A6>1qJP9aJh^Q3cj4}N31FtTF-kh0r zF3grI#>4_OMr{+3ts5XoAFh`e@pc7>`(C@!in+>jlPXE#jyxwr@Tuo0j{wK6?q0#w z6K7`zinb5f{Da~2$KTL`BY7NeW8z5p8O+DQ#@}K@Xz*%C1pq;{5~xWpfrv6>VTnC3 zr&ChR*oKlnefSSi_e9ttFK&bG7i^h;Z^#Xo@2arFcf^WVv~Gb@;jBV}_<-DfMs+on z@>lxgR6w`Pz%(1CU(FjCM*c?GqABd9l{19S=3bkzyfm^m zKhUo?Bb=5KjWALQz2U2nhi$4H)+HwU!#QaM|3L91~>)_)O@6UD{ zGVB@WE zWCNE1$AaeUqF;gqDAII+^$Pyv-&>gW=fz>}J=RGFD-}W}Dw)qJSjOxC`QOhkVm}NJ z0;<{8{AJn`PvUaul#ZH~#5(LqSO%g`AQBcEG#TNxD7uifkVT-S6X^x__3HWD6(6vu zqx*Ejt0+m&v-pS$-hiN-99_pjC>2X}_cl?^ZiP;{CEHh~?gaSQe7tOoC#})=*Tg^| zf|&V-t?|i3qm~u-SPB;QRA>r8FclRd1FIvrN(O}ZXo|>tU?vTS2*&aZrniY+T?KKEXM;TQ{%R_j;Nx?ET`Ok1piq5Y4%a$s230l&(?v-j7uFCwiT zs;fRx0v%CKQ~@%X)E*3yh67T_;JnvvmzXl7~Wmbk3o z?qDnKXrnurxex!0pvBn2{mGKxaLfO|zb$BY0Y=0EHk`5VE=o&5Nw9~rTaFECXc^N6 zboE<@OAu3);Fw{@9!#W)rh4DOnPthm)YtVMYf4JCNC0Sdd9h}XL0X@ba1V_NZbEY zS9i+{n@puUuxK$G)eDu`nT0KclWJAzFJmxH#l%!sA;$L*uf5g0+3Mb8FV|w^;NN%$Kl{|xAw|| zr9~godlWi3z~pf5-3%`jWkY^QL74jn|VI!D;kiu9EvBKBJqs-h0rU5US%CM*S&d0C|Nne zLr+gPyhc(L)FcRzNk{OT3%0_=-W&a2$B5S5=~2seZz zT=olt?IK7kIngMoY zqk#g3tWl&O|iEK;nMY-&ccO56q1b_g-Mj%y^^gwTEKc68hB z@gY0X@2@dtqCQ*T5N)HTI!YzW2I%DQSiujY?QmZcH791U3FNu1-}YM1BxT17zPd_{ zA2UC;h|8?o)zbe#8{NQc&$XBDSuAitGU=)D+&K>kZG2oFySst!6MOSthXE2L&|y@? za+1NOMNvB{ExYntiR+@$!8*PJU1qnq!GJ|fEUr5ISp6Ty!G8e6A4?jLP)| zu%PSVnvB#P!9f)5dBwy(&;UZC9*?CSHe*7(M_10D&hqN`7t&M7O5#!57?`|i6KTJ2 zZ71CJgj?U&HtzEm!+(D9%!(bbum=jhSV;(&?_MAhMDIH>2bJUR;vQj7w6*?p(BO&O zV{rX#ge%$paRd@kK5zDo{du$(oB;mFuT1Vz{Lh7IjPlJSffkagMA&f|F}jwDEClk_T86>R$wd8)SH^3+y{g&| zA0}_twkpog)Nws1+&%=uNl-t1*A?Aoy)$K7y@t0<^=klaEdwwbnjZ4R==1b9-%<&~ zVOAM?>Baw%cq7J79bv&*^vwo&oloFf&M}-RH`DBM>>i(r)mbZGG>-Xd8#k0eR8k|u zzk(5x8i&P?V{a<9)hs=^&BEnP@O^%Ph()HvF(qZ>ql2qAKN%Pfpv$TFhQ{2g+%G}I ztPdSN=%0(B^8l_!WL}?~#MH_o^uZsNxtcdLP1n!uZH?o_HS6Ys7Zypebi=xsJ`2?_ z%Ca!&Cj^5dX7V31>EvM(V((E8{s|*3``M@OrhV6});M}*L6}9=2&p*Xj=UWZB=@Cy zR{`BkQv=R8P>-&X!#e4+gTY`}Uf$Yp(&=5y7gh+?Nx)n#77??b!HfG=nbDts;4Cmv zKaq^OAGATNON5cD6SnajI0-oDOE_M<$@((c8=Kf90*;CK77qHX z(-G|`n_p3>6RsH;DFVlM6Vq8<=GblOyS1=k$(UoK72Tp*2Qrb$`%v0FJBkHyxiy(d zySs>@oPe)I6NV~w$2?7M>dO(GSH=i&U}SyXzsii`W$!UQbgZr~>RTG+ZI6^%0U-5&h3A97dA*;eHZ;(#Y6z=W0BxU+%KH{5Zqu#&`b7 z6$cEk_`e(!eI8FRmSbES?e9(lA2KWwG=VfhDf-5}57Q!uVy z%Lpc<1HU(74jFn`^9+xE&wpqsy4v_Ddf}?OA8HNd-#g;(c54jm9R^|4MQ2aMF9Kou zFpOUC4lyI;Br!d8&}pbY;#3gvh^uh8@fpyD%WwH*$VMZ5OhL!Owe@Q@9Ks1LdBTXSz`=*ff0 zjx*EiR}+6dF!bzlCGx6f80g?h*8mAM*mXD~EK4!y^;!r@*!2gPpzA{kBVW5v;;@A1 zDdwnI6SjB1G+`RgGQ+?LN)#ZUV@Oo(*EMD=U@$|`j}4_(@POUxLU&^+JAQPT z(}~YH2QYN`)W!Cn#CLo%i+Cf9*`DcI=JC>;XYvSFq43;2UL8k35ujJEtGC}#%n@3k z0-J}j5fIfTvelrp6N@1Tm6v4py@cY2z$#mbp*9UB^%2;hPee(o41XwfAm+e5qvh5u z;>xnyFcoJi?y>aBR!k%(7g~!}%1?M0_|mqMDXMy{@L|Aa^z~Ku>eiC;<)m896em^R zA^pQ|qpb4e^p^ba4w!4sLVj;a6fW=I1XSDBIB3DWE>a8;kh2#soRE~+KvE>d{}$5T z{)Hp1kDT}+L|-(FV!n!`=j1Uq&mp~Ue{Ay1Gma*0WrQj`Me7w!1_v7P>WE7+ZBL4b zeuPLJc72~6Y?Ji%I>M!%ggZA5ZRWvUnJ1-_8+tU<-RjeNWd*+}xCn0&1)$?EjNdu^8HTy4!;+2p+l1=^_S8wV?tzCdtC$>|WVubH&i&hO)53Ky zl2svje&SzXC-Gcv!0;`{wJ-fY2UQ{g8gsf>b;H`Qz)fq8b~~aO79uk+vGr91S(%mG z5vJ@=9%qJ~cHIcB;;F)Wzn_2%LFM5k)Vr98TM}c!CD3p#oKE@Jdf@7^KN|GBP7EVJ zw5tfc(QS}UTCZ9BXz>Vc9PMq#mwV}cR-9H;P6M(2(|4}Hk7b4iJ(SMP47-$!iXNI1 zlQv6WJI=^k%-rbMBU^nEOPl9hZ~y9MKljishx56)76ju}FQ%W_`#f;#^xdAy?H{NU ze1$B62XG;Dqf?rya%@}~9gh%kbbv@prCRF+MW-V!K^dRb4Y6z)uc?Uz!>?c8aOW7m z_7L(3dPjpQ%d!Sjf5YdYw6-?N^r8ai{iBBX^u*`-a6fXTts|GdfZHqcoI7M#AQ7A9aOhll&|X5^h>^U9FAX)1i*N=?2X7ZMZa6;x%42w`BgC8@YI^o z+W=0v#JVg${viqZo9#CA2An;XpFw?ap+iK(e9~{=jh&>HINsqY>V}Yz_DS-rvIG~{ zRo~CxwbUH1_5LZoyq~|x!k$+0*z3%6>uRsc9M>E9QQ^FEG+Ks*X2Z_lY=YB(gXrdv z1T^o9qcG7qm_4FszFq$MQ79H=OKu8QWKf&s5c-tZXXi)vS8RQ4gV!Tq z%E*>@e4ZtI&wh=UR>mzym^xc-%#87@O0a}y;d#^^2df?c7X5rBl> z^mpQtnXI0x-+T$XRr%bsZ{$Gnwx|*a7}W4ul7PhjkGk6!foV9 zMBBw8Fq*j+CzoTEN(yPGaj`1X>6%$;Sqq4teqlk|G8*WLniew3b!>%kpXSV-2eMDl z$ANbo6bgQ9spewC1hOrB+wDYm9UX*%F>eRHmcgJFzD(OuSl1XIK@AV*Cjk7GEy|?z zLI;gk9J9@{Dmy=c79QC1LGKEZQ!6vu#$r^xs?pRIScVXmQ>q^;!HrZ|nRNI7RLjhO&*B&C==^s3Y9HZJ{^FF8(syL7!t zRxWSsvmo5dE$uhfdaj>N*%e3jF4w`A6@$AsV>tmmxL7m+f%4$qR=|H&H#o$fsn$&m ziZOp5@Sbpf5sM4mg^fSa<*S2@4}yUHO1DGXR)NN73g-Jxe>c^4hU?8u5*f!kG)*{_ z+YBuwt{6S(Quu}vnLOm~VncEL^MO0^Mf;3gM5*T1Ie*Nme@mKyuGFwD^c1PnZZ(=C zNKI07zPlvN{V9GXWB^PDL)F;p_`qABi>e1K>*HHjhL?lV*Bu|Bazzr=1vr1P?iQnw zQIw0VR2I|5#cpHUtB8=jSYdWTjjIny*SW(t?r_F8Z8D?QxjoJvhOAw&-8)sh%=fhE z8FpEN*|kKn2FZAx%@bQHy5oIf-MA_J9RT$5mDewMtbNUwiu zw@0P>X@?WYA0-OTsD_36maHhpM0Slc=e{DnyJj?_Xg&~belg8<IKPwXQY4>&38SHPd=yymXY2>aW8>ONTevT%&qtT51-hVB7@d z*x@&8R}!981=mizc;h79(U=qvkI~)v69>k!SSid4gfsS7V$)0+ znVZb{q3QJ6Qk2rH^}rr~2#8XHr3KLukrC&RQkLyM8;w8)VXI!~T(COj=#?TUs@3#- z34G#=-G{A#p8+-umJod?r0{fmyqw7HD`YrFU|@EW?mOMnYamw0W|qTXU3R-(Jd5qd zGfEqiBsgb9*!A|}P5S}K+TCNo4qClXvYytoxzb3c3CCzI%@npKfdUKgI!<47|z_n8-G8q4-K6NnzGlI3s z=$PmJn6bo&DK(#6e5hhVS`laQE=mpw>RXU=UD-o`SZPiq<4WF-8Z-DF+RFE**Hab6 zVvozxi5Tqh*mP=G(0u}Ycf!6>ToZP5?N%BR8Xd3=aPH&0@u|Urs0pixm99t<^lLKTc{vqp(WGeGZM~H37wimj8^@?+4*MIHZu@ zr@#+mC7NQ8nduwr5U3;9%;ewFT z&U@*|(2HtYZ*n_xy=RwJ`1f53PcFJy!a*h3U$7IG291o9_Bwo4?i@OObgXALJTMof zVqP;>`ducFPAkBpjMdJ%t;7y`@}J0-Bmmf$Q3*o;S$fMr>|uTnScWJ>2LJPH+>Ffn zouYczyJlD^(9&-3?;)gNrBwRTiPwzEV$y{ojJFSQw99@Wx7Y)|Fg_<}K*$G71o`ND z4-YPO*JqhwhAf;zlX0BCAQ-swsSP^yZ zsal9Vi?=uO^Z{8-I^@UVfw_78i3KQuX8jCGdoeyuua@m_Q9?N-^Xo;(QBR!Z?$chU zI~Z18WAeVw79z!trGKgl zS#T$QD!UU)h;fI_LSU-%3apL;n$FSm!n4NqAV0up84r)WM}n0!Qj82u-N)1(m6g2G z8}E`_o+rYbz}*-5F%*Vyv^zPUlb-Sn?t|&Z$@cFPJ#ua%@cnq6uyAgOsQ6Le zPv6*uH`V|u*{IpS9dOFUfV(i>pwDhuLct1T?e?L%=T1?XB<3PCralTTG zvSYmSQ$<>Ik>iI)l?B1*Cna|$o@vVyZMjc`htm-KX4k|>C3jx&y51{i>1g`Byio0w z{%axm$KkD^m84$@0EUmYU%17Lwqd# zouZ%BTA`R*pSMV+W0=F?l|m*i7Q$a)F?K$%rA$Y{(n~;c53vc95OrT&jGmbgK@n%$ z(Y>O)%7HhxAsK=lZI|*^y2x(+;Q~&-D_#B@A^KTfC^W&j;Fml)XnD=Gs^!JndIPrZXVoS0%d zD7z{W+R42A7E1`F_vxLfIF$RS_>MlPpkWN8B7G9)dq~5yl644}w@Bru=E1mHn zl83|U3Vx8>eg=I#?x0Vwf+ftr^z9;^sDkB|sACP8v^xBa6WzEW20sHr_iwU&rOqxU znDJ|mkZ*M{vxHNcjmUx+4FhL?G@8|q+o(eD$a+u{%%H>T+`c(R>)26kRP5Gls2bz2 zs!u#-pgX@vw!?09Xlh{uRX@(5nQr+$A=e{Nl1fCezvtBi4UZHFt}W~RhNnMBUsgi- z<7MB(t5k^gwb)dnIX9rErWt2B&q(9&7>WO4M?K9x?Ae4}pnxQ}aXDGI1?7JAsj+nR zmywAMxd=#FkV%0?ha;S(Lb(@QAy%UNO2 z(Z$^L&Y07MUK-^0caxhWB6J3)`c2}@(|V%Ij7}`{?U3D1_@^2}Ut{_Nw(bpTjW>#_ zN!rW`uVfR~5`j14*#~&E<37a?R?U#XDKst#oNEMr@i|6Q4c5cG9avEOyLUM>@t*ri zAn!YIsx|pzwt2Sj{Ez4L6K;Y2E< z`4GwE%#g-ffZQuYTTJ$^50bgMji@SoOyndb&;_EDYe%Yzs$FVz2R=nGpBMtQ|%uavXiw zZ@5UJvz5q3w~D+bPb_AcH}0}~B&}5)?ppZkZ+8Ng+lUg~%|v2>EoK@|^~3;D2l_70 z<@|Eg&v4_d8g@Mp(Ass@1Xl~+(C>aVs(M#ozs!koGE~>BEJMEs)5P4ZI#C$VO&Q18 z8U2f;)mct)-MVBBVOOs-UO9op5iZai89_CnF>`6!!Wo>=gU zm>h2I;$ z{Fq*GA>>Xe1mxFFw?(##PJbL?8IN(hWHhoyk1=qe8+F|%MpV@=c0Rc^cMxaZ5?yEu z?*F79iDu&2!aqwxY~zn!TYsk}0^o%-lxobjvr;BuyKYK0Zq9o>Ft~wW;>!Fa{AP0u z554{`zCz|X^M}7NUfOhet`#Rn$%~x~N4_3W#9K$rBCr{*ZOygL6`rd4rKo$qlhDXA zG3KJ9#OW+^s+uUp8=6Pt9TiRR#GU@ghFZ?B_pkWvW`?XsDcfc^ut>Rq^Se`ik=H8u zOV7rwlTWnxh^xw{n6Lnqi2nyzK&QVTh!Q@sj)3Hop-cnb(285YGTxy!1Ot?+bhSRY9usfr*)DOBlIkmEspQ&llae zIyae4#}}9+WPvgv71X!c(-8ocpjVX6h7RU{CmW~oen9`p8LN@2vKx3ue;_YLnXr}2 z={AV~rlZb_y(b!hD-&l)>$m!3F)IWSn7oOKalc;|_iHTR7&&Kl;P zRqF;Zz>s@_%bKLXHErqgT(^5l4Aj9L48|XC{X>w3w}ow*AD! zI~SJC{PBb$20-Qic;m-0Qyv3TJ7S{;4mTvWa1CGeo2LKLdVcFk-c%WVf#qQ=Ip=Vv zA5S3AIAE(_AwW{*wELQSDw)ChN%@D@lr^uoWN!V;Xml;U-Vj+(f9EN=FcUJvrk0hP zUh4nx*WIhU4mt0h>!`kq!;-kVd~6F#S;)3RJ;fK6qAWS`Q^%{cXV-W~PIpTwi$KNxN^YSB21xH~9u z!0gf*6&Qf*A-lVce}Xzd5Zatyx+H$YQ#z6j_gJnzsjgN}4&W8pkUFjDH46r<2YIrI zw-T0WwSvj>txip_ePS6Vw9+p=A!x-%JU(*0i(J$bhA#ssY`#N~mDs z!EKwlAkcmv6mNAYl*xtjaf`%!2}xpbm7T+PZvS%s#LiDSe^idS);panWuF|)S!U*2 zgOO56B8?3t$32|32CJoB!@`FNo@SLog|brpdQpbJ5ZULp;H#t2LDnj!d%2CaA8l;j zi(qmjgOXjWzA~tG?_F9Q8+XvNn%W)p$%ftIF*3I4!FdWjlODwP+XEcRs( zZmtSSRZCx&e@&{Y$sLuN)huEbH+bNTlk4rv%lX`@%y;BhiBh3OHj0wCi6x7@>2Kaq zhQ4O}^mj4OrDt*3AKOLW+7-c%=8+9zPIRIS5zso*Ra_)H7X>nE3kf}QuNT6D(L*o{ z#xyAz>Nw4pDRUKuMJ>>x%>%s>mbh9RvpVkmowmU(f3|ladktoOSkCb!P$tjM&Y!xS zGZjAYc?jnkEL!pOFW0^JFUV%hhIg@dl%CtJd#giRDrx^8|I_mbxlqEn-Odl;0$*A< zq7uPY31RuQi+y-I@mPm6-4n)1o88Gfb*p&+Qs9&~9|w5@Qg;YL+5fAJA^4z<16^}x zXZ!`de>RIL5e{j`S*~UZo>m6T0kN8c=498h3DggKG5W-_fy}S?UOquKqxPz!*fO>| z7+ukLQ{@hjdU3=j$3dhV6M)ZzI5FpJhNC>=++OG(M*`<+x?R=#jE^<(5yVGu04FyT zee}O%Cj!77%j*+55sOg|8f!(UzRF)+MW8*%5DkO%@`#IlfZ9h3tvjs2vKp zvI#saB3=A05n1Kp{Cg}F=F%$TRg~G53;{daYdv>Gw=AU&ZJ1@E-COgZ&-`=DtkP`G zTz7p6|2L|VCro`W=`C&uW3q|!GuH)d{>JtodH9Jf9*N}iCfS9oRfw2KTBDRNVN+ZC zf8-`r6ez`0ye?Xyjmci!S90v7?j8(8jtJ&xOUIF!cUXlzs(TKSTY5<(XHKi1f{ps9 z``V;l;(fGn#v?fWmOfdz3lE)k+Z^YB&J^_rkev-7Y#OjNLF^7;oX;441^f=TDHd>ExhXiOD+~Mq8okFOM0%VBZE0t4&ME{n#@X9_DlcFfU^Z^F@fU)VP~ zl7wV?%?>WDq8Qa-tytUtUjW!Ae<#@W2j`%I<;#p>9<4bk$}?FqE+wVu|56-w)m`Fp z#{AZ^@AfAkHZZR%*q7Tpec9CX`)rWJDB4kiGgx)$i$GRn-J=2q{>`>&DAX-uc0y1) zkov(c&g${+D4m)*ktNb%IPY0*KLQiM*4N%-c03&|@^S9)=1cv;TNicLO*-l05*tj{WEbQ$-UOp05{>1si+VGF=5H zRNtqYT%W4?{PcpGT=jd3yZ(*gau7;c6HHglT+lFEzlU9Lhs9=vfd3lCe|dwl<-t6v!&T{( zsbblZab`}ij$TyE`Tpxt7bDv!ylT0zd`K#jCx$mL3l}?X1-=$NF0CXJ{xyJw*zo7a zWK}V@3IQLv(29F`P^5_wNMD;bNx#-p*fT|HG?M1WdFk5t+Kg&yfJsjRr%eKH=_9e> zc+qTaEJv4G!El01f1;Fz`0zOl-y*!MtZOJn?jQv@^*@6@2Pc0^o15Yyn%cVi*Kyn_ zuc@@d0t6k)WtJsS2jX#h{wzMAw;Fn~kG@h>emp#!OlSx#VxH+D6gZM}&6#dJ_q|Yw zs4nx$iAN{i;p85(AF51i-87didpA)|j(5+t0)G6!J!wDHe-Jlk-S%+ryV*FD(~9Km z9$PaqmFzD}q`?<4Zk_PxGQ&10e4jsp7lNmFkZ?RsUEH!G=Dl}yHw{~Px=0~3<8*JG zKvY(E)4;Oam-P+}P^%K>uYnX&dYJf;t|dJpeUlYZp)3X{}^okW+<*$Gn|Xzug{}@u4pTjqFM}e7Mf{}(f7;2J^@ht(# z=mK*sOCzla=1kD)P>r~gVR$pkBC}l&zT|!469C|bwzn!uY<0CBPM^t6#U2gFNLj_h^ShcZ<$wQ zEN6R^->?I@(%V2uag+4q>vg=@nG{(+Kj5Ksob%~#7CgZ34uh?j^namXWmq)?JCl4E z??>|kKS0lnZExKmSHAan|MGKF4VW8nbP4J?-d_tsfdgnbhm|) z>|E~Se?(nLyl_@~n7iDhxA};AIAZ?|_%gb{Z9XHzQnUo(aDs8nN|^nnVCX2Xl46J= zEW2NX5T|>yxcy0)q0LpkT_2JEx=O8aFR}8fjF?t40cp#KUuKumfb}AF!o*Y$kqQ8a zxIpH{4*PSAQnj555LqJs`3kHNwW^yOqj|QJe?=WytDp(m(Cb0S{s`u;Aoqo^@pdIz z9)v?FyAS^(Apo5rSM|ekcy#cARw_+^4kpme?)=cOX^hS(Z*#Eh3p^p;~NnU>@n49)8h7hjTjS9%7Qf{F{;`xIbj~qOW(GbCr ze*hK=-#jEfW$#dwuLZT;C>#cJ%01vZ4rTBz%Z5oeg?5LuAy-cU%GX@AHp}D+DMB^z zZFXEfWm~9!Ub50y&p*eGHq=DD%(E?7D*);iBrGh( zzt5)KLtc;r`)uZ2SZOhj=0M(TjBB8fNzeM2d~ES#!d6SsP%JP){DuuGSJ8^^f0dG! zQ&Bs@Ix6ex;8wdf6D>KUL2ZX~pP9u+6+@h}Wl;Cwc+^Qb>_VqjU~Zk8Dy=?|v!p&R zd5Yp0u+YK^0PgJpVsavT1&ke2oS8ksN@qAIS>fd@OY5vQ1)4VKqKqID^})dgqSLl2 z;MlmWk<+{>^M^&I-mN749peCHf7hDf`$B{tk%fTs>O<0mtv~t^l}(!FYu&~(?+U+V z8`D||b8E`Vqq%uqNj?_(r%{IIgYL=qVbjK2;?C%cZCb&A;|;ctoHSS-GQo1_%ccauc^uEv|YlyOz>{uM5Q% zu3c=z;dWF?M~0eY<~^>6Q-5lzS;fzJRMZHjcqG=|C$K;NB=GK!Vj0kgE!k}PxMe*< zzD!Ka2^!QeE4@R@fdh%2f2dJ@TXl9|QGa*;DrdXen!_ zAG{E`)RP^7+?On+r*3YXlPMa&u=rDczBDZfOSq~%JgPLpV&xJ;g#*kS{}kgWin55=Pge-ToGGIXX@bbOc? z$>$=s3c?O5xHvugU8kj~bO-a^^g_7Y&Fn4Na=9fbVE~cXsqm3uq#LYi6>@y(6s8W% z*js&RgBCYymD9SUid_x0VksHT0qgm=IZ9->2N|VjoxH6i9gb@mbX`*hOFFa4NtRfA5tM_URQPDslRbKDaSV z%eOnmEEbs7BF7?y+!>;7Q_RDu*eG6QA6lVjgBF#+KGhf(*AzVWS(c3J%fd}!J$ z)Lg=TSXt^;F4L@2y0o8an_P*5Q}z&*fXnE3<2a%VkytH_39i=0-afMT$&AjU;I}K2 zUtE!~vrRU!e-jxs=%y>-?|S8p&xJNt_FUS4gT$JMNLmi^G@fq%VAt`t zjXYxTow0ZrrSYv5!+k!p!AHV`#6Ikk+++qgB6v|4lc;0|p)KV5x{`th8LrjKVh$;%9>ke-$4XzjfyuTW)Y7xtgV1loJekO8Z9f1uEMi zBKwFU^Qg~300;hMA@8mH@u;^+_yYdQSk~j35Pr^zKE8frU&RtbB_}3a1bxehU26p2 zi66)Zf8EFMqp>PXXveS;h$_C|Br{*EpWf5H!Q`(gYdd4vGvh=w0&5l6bbC`SPnY2q z_01y79P)s#)Nq_*tU(1cgwyo%7!7ZI)*h0a;O&8YsPnY~t%Z z*n}qgBI<{=^NbOhH2Y=84CWJF}#T`VE1LMV#IRT ze>oy@w=0l7n2JW6a#0y9xb-}-1RdH1o4g#+Fj_mLvl}VguO?scN%?8!GjVr9BvPq$ zfm=h9a=kBik@Qkh(@97c!bId>l_*N-Ce%s!Uw#W`oDE;#e=ZF9szqXmM5RJ;ir;NR>_05_(8*-WKh53Z zWZ+K2vOZ&@4q`fw?6W+s!m8gIZLz+o~g za@6V3Z?gs-L`B_Oh>&Yp-{=H+Ub}%bAxF6^Nb~+fOaONr%%J7Ny7$E4&*L)>;wtX_ zFfer3?(>6LYOqmq3Dl|J&UuyLf8x9zf!?(Vw7w+dzQvB>36}LJXU55g9P*-U!3Dg} z63-3|x-F)eX)QmqfFqeLCbrKuE8YWIWlb{l4E`CZx*h!nZN@&E1~(w|oNpf^^* zFBp_)5=4{j51JgFdO@E-8>_j#&^Cl$bqaNCS&88iK$J)GnPAcPer&^y?Yk|wzIS+&Yjibb9uNqyUXL-dHLRoz61~?cfOve#8n^g) z4Nss?;$!L8^f5SLwAd=Zh;^pl8h&^){?^z|Z)H?uYt@=P@D)A{e;qeUNMozchDkQs z$%_ZU3JM%w#;lH5yzY2xD9uBO|8`+w&eVz6yb&!dQP|l5BPVr*e%`$Rda2>QBS}c-GFKGK zoh>skt^IOA@`cqAq!6Rc#<0<>_@AryH{>D&*UCX2a;ic=tN~hJI9$OO1a`4|&7@pp z3|K!F+8}T!5AF%6+8Pu~WLL0-^1190V3HPWkalfbr*t@!&YPb@?)@Lk`5qk@W2~ur zvM*Y|`y84}e>^YQY9pGHx#-3IbzDr0js0SiPEMmRxTJwp)DTE5p&%zVq6k2!A&$K| zXSEvWu`XK<@@Y%$IbHy^4y-4jqbwSfo249NEEpe_^NP1Gnirp51%Yi>l!WtS4zaNp0Aq@T&#vr!8%+EIeRi zGWE8>r@laD;psu92+^zVh%bwddB$CLcmcac*7FquJ`beer0;1SA*b|q)SsK4AvL)w zuOpez~ijeA`Qmg%Nc#0B7;2NZtTe7kAwpvnf%<8V@b` zIcCk1&O03Us%v^YJO6K== ztrB!S%NKve#xJ%q=a%zzb!TBK)l&)%d1blkt& zrIB);5@*=gxG~5eKh;MAz8=Vz?k1~r>Y99CFykP_6jPf}N-?iJeui^zTLH~ie-*|< zBvS0`1dYuUflf!**kY5ZD$n~jLPv5+IRX7tZAM#4NkIq53N|+Lx2!kx75y)cr!@MM zl8M^JPP*Zp&WV-6Mk~p&@f;u_pSw9O^kMn2L4{hUM;p9)wMzP`tPG=s1YN>8)bNAX zj5`msmB8R@ruES`KFX`#i*#*Ee{Cb=<~H5eNhoW8rNSh1w4cSmlc*#L7Rua|P{KV*-;9QecJy;mY@jWQ}yNc*>^wx}_mtXzWwhzG&XmkYNo3V-LIjvE8zZ+D-*$#Ua^b&X*nXtzy*=Sg~gE*2-e zgv-Y_(+-}X)P$L0v4Xp$n;vD)9nL+K_w)V#7K7c^iD~b)$S8MSuG(|bQkH2Swk0Rp%q$e^lH&P) z$c~TQr*!Mv=c!gO%XPzd)5XUod+*^<#VTMzady`N@u&AiyM%z3Eb!hzZH6*Ib&U;P zsR4}Hn*u`HVR1 z%DXxCA+HYTvQC50ic=i&6o3Nxrr#NVhslY!K1Qo1)`S=d)$p|~S>~^1Ed8mh7^66v zGbA29FGsRv>`z>6J%1_{v?O7*MK%2eF3I3uup_4AO6=zEMl3OY?BOZDFlbNn`Yo<_ z6uqmMMQTslk(37Rl7HfD9DhcZp`P(|Ax^WkJywJ0H*~?0Uimz&L*T1zP4l=C2IrY# z&;|ftRM3QRYeq-x7|RQ)(r36qXGSk+PRpj13ZGUDR27u1!Y(rAH>&1}02` z=&sk~tbsq-_6_^gvA5s^9`=g0ubY%_`8HsDTtMB6$`@Q~Kz}_*XA;(zj3@GdAGot& zZk0P&Sbde0ulvX5z@|u;&74Q0lXU{`qZ~`6W~E zI?kLH=*pyx7~XE=dsknKROhg*QoEW2aF;8*DC4m}%6}&yzGw6QC@5QEd*dJGFkkwB zP-O&4ozlhW;9C6=7Qi^L3*z1vcjK<>nG&{2RD7|4;fLJs@FoQ)0_O|MEMvTvSv2?? zxMjGi2cXUTcp`wfHZ<{(Yn#j1#!g2v1oWF%w(=!oSvTx|A@8rUq&(uYl7VA+JTyYM2V1>sf2UUPnsx-cXiF($bSY)A=?jTjZ^-oi zvc-Y|MFVw6N8*iu)X`9N^I#*Ep|``)kZvd^tA8)p3s;>1yPDg=L9?<3S`}ci!r8j& zT-8FA#pD=?)NrU^(gm(J4m&K(J3%sFk`e!SV~wyO3$v#+_?xO~S$&1N!6_u-C1`*1pb^N*bDEU>@& zLu|OKp5JGUPCiIrd$X1U&m&`dTc+7}9lPrenSr+>^*6ssT~X6_pU?H{+Y^vgBj2rm&tx_`D4 z@C@_X0=n3r1)?aETNtDX&MTVO4Ocwqe!h z)T#x+0M0NzyvIAW^iTM0=g9}u?0;kSLl_054)`~SMG$IT^b8X$H!Q1ARSSyb!j z-{k%m(w6Mp`R`_q8fw1lIG!!=-#0(ytsi}zYUSz0J!f5e=z09$vI?BLIZvH96`X`~ zsIVKMN8z1lQdoH;i*y-w03f#Upg@Ec9`F$Z-1BXTO`o#=)GdVz(;ACRcYkFb<0~X3 zRUUr>)dcc;`Hzs>$6vQ=y++<1SK+rft@iy%=A}FIjBF}%1j;HmEb(&GYipE2E`0d1 z#YBnb)~J?CyXa}thu4x6y={uB(d# zfm!OFWj}YccorP5_A8(TKYv6dN};AntI|BGtG337_Y{2r6m{QD40-8iY)0P$O4%7G z>u|ku%0*8^SkrTks4p*-gNoH%L{eK}sYe`h_S_Vu3e&J7l8x8bhD#^r7G%gf?#*aN zqGDzW8CE5be%PokN6qxKuHZWTdTGPnGneJ$`r8QMT~pv*8>M44;D2^kXstKyP4;Dg zI8+%9osgDsmRiV!n>&bGeX4M+)p!FlYA@n%uJ`vSfsmW_T(FtCZ7Y7*W|lMHTZyaFGbj<` zow-$RbvZPCvt9;9Gk+4)g@o!An6!S7CXE+6_*+~eW-GiP10!6?Zi6Z(*3rUxjUN^noMU_cfHeV#>=!EFdFl8{%7O}8d?s_m9u%N;?FjP%Q{V`F#iO9-w zs#DS;*i5`E$1J|ajTJ1sh&wNa@xjl`<>7AS{?f4)3=#VpulP5U%?MTOU07{MJ*JRm z{U{ok1xNZ+bbqdf^!1NY!);rL?Dw07yhXPzqvoY>LsAu@N!tcM=vvCs?K7!by*R}a z%yn%0*?6G$xO=ReIs_)eoT>D_&fA{N&G|k#OSCEK#yriB5F6p*L@q_G@ZM8d{%G^+ zA6?Td<`_CfBD~uS?WL1l{T21WfdH3<))o(3e{#tV41Y9qkkz@R(LG?ujjWPu2hF(< zyIGVAtU%@_)#1mabh;Vyufx`G|J~lP@gl`lhP)0OYsoTl^)hWSLVkq*Djao=VZaxOS5_XCS6 zmZkS}1AoGpOGQb%IEt>q258VJI;yHJJk$UjDY3KI@_0>X=4vs6aW!U#8+S zbNeev)gQPB?v^BC1Zb42rGRz<^&Er(vY8u^TD7W|ZjpebuElR9I^ zj3<-m?QwNyWuD$c3 z(SQ4gO-xQf&Hbw2&I;3Wq73Ta^ju~fvL%C42*#Ai3oEZ^nQIU=-^NzF*P5W0FbFF=3a^n`JbL-~raaMDA@p9PT&9*s47^w)wOE zIs-0bV{do#pw1c{Qqw8omA?PAVEduErE_P$!?;KJ7$kgJE+C16-;r*D7kE2SNq_%Z z2VwmK6Tvr6dK=`6ziBKmtx?hf*O!-iSKjP{Xt>SIJw*6Z900Df7QUX*t<7ubEycKU zQ&K!C>kTc1OKc7=Ozb$yOv_&;(RdJR_dmnmsNpB!t&R7fYc}DNtH!f1uuzZ+O5BDE zlTLCY7VI1!3+qC4X4&NMg|@5?MB z%T48iJ{#P0GBwYTlmY2Unz2@(mTm z{1_s??Tq=Wjkgj5LPh(?3x9=^Kpk8sPL;oq7oZ{i96pF??f_gRL}nh{U;#TjInfov z!-I~E9M@v&XfiA>enm;7W){gQin#K+GNKgR@iL(Qlsr~yv7V9-5;Uwh9uqv!a<>*o zX4<+6G)=@zY_{ukf}qqg6JrqvE<#a|81!Z;3)W{rJ2e7M>yH_T+<#o3d{ZDNX;5xB zx1AHcr5UqI+eXjGnY| z?mQ}fzs68~%AJOf_yjJb7lB=r`FI$>W~-#ttc3qI7yE~emrBq|cL~HqBIyK3_~lHC zx)Dt*0>+%c8`wfc5U+Py?;1Nk9AK*E8VYw^cwG9 zDY*NaI`OshC0O*JdeI{FAX(f>QJjk2C8dCT5JvtO2^j463NX^)W!B)S?v6WEL)u_R z<08&dyQj3#hF*nFI6}JT`JxM~oYH6=$+n;N;wAMgdag%oxMJ4NarEPoa6kMEjbLh6pj?De&9fdqDk;UK;QJlsjtZ--myl40&QK$q6wEb9;oV=y<9!H>~LRhHRgMi=DFfE zBoX=tj(>YfSe}Ar{C#UNM;o}W4s&H_I>L9#p1;;S;;f%H89C=p zVovuCj)#W-HOpA;vaJr8j6jc{H1)rd8h->UHfUWECuV>W*tU|XZ&$1|gz<=lq)0lm zelDh^Ce5|w<0Uh<4IbLZGJtDP*U3+1IUBY5Y=5wc)!GW(*$`|<4pqInSR}tdBx)F_ z+2Op0tttR$lSiqjfww&<;9O#}JmPLtsQ|0B8U=`mJnqSsm6Pjw`Nd5s#ja4TgJX$s zxl=QZ4)yP!&!MX{rn@0xU7wJ=mq#g|Q{F5%JG8X#+|v^^2P(XZj(cg(C`(W5f*b+S z`+qMuG!zK$UiW(3(m!8^xVkXhG*w@@XyMNGlOXEwxk+*(+bPcAeg@xzVf~)LD%u

                                          ;PIuL&eTfY@eTpABmds76uLOPH=Z8xDW2` z?hrJ%YjAgWcMa~Y!JUWv_u5T%&;9l}92h>;RbAh%uI}HQIYUY)r%1zZXsr(vv9@xc zVW6ew1V~9MSX=5@F)-2ySX&qZ7-{M0nPEvu1?_-(4yM*tLV6BBP5_&O2|(7s;XTj} zf51pj&jCvc5CvKR?cN6s0s5{0X`q9ilB*4n0YI+z56D^DJJ9Is*}o41t&B~rfE4c` zg4Q;!cBaNA4!`GMqM`Xc)9+{jT7ZO}fw{G_y}2ns&&m)WK`TuQkg;}tH<|*-t*rq1 zKodO+BY?FLKnbV{P*xOHPy~o7$STVze^SuCPps%@V`FXie_8|;m6S!P0YdyTO2Po3 z3N=7fSyActUnQW`yMAM8fQ-`n`0qULj=#gDg_Zb~)Z~O2=zgyOzyNRp+S!}_PWwOF z$ljF!{z!WdHL|m|{3ii`+{D4bhLeuY+1Z)a*wNmB*4oaP*2dx=eM%;#_5f#Ve>-!) z`>P$$0{Bm19IXuB%jsYO{MQ7(w*(+%Y5=sd2mTHcvHsVf<$EdLgWl~9{}J|H2#4Qu zTKvl$U=IZTR~i#N`+s7kjTe^Vt`^q#ef80Iwod0*n^sF51-Ttc0e=pm>+REP4-ogG~GXeod zrWU~8>g|8;nW@!3HfeqtaS>rfB^s&s#6G7QhTrD*5_r1XSla-M^epUw zo~A~?_ZL`qdp##0z`@QD=;{9Vz<(1g0~^55)WG4r$=?qO>_4-MTNzmcIR0gRSMu+n z|J#1@e-0_d`w2C)wz6;qe;5LdVCiJ69p0OP{QsXb_CL8q94#zl^elno|6S1k_0qF6 zwQ&7!zyERh3j8gV{QnVbYA<5y0yLB}bucjbS5^OI7kAKmZ+dj{IX%{(X)t z-W&G)(lPygX8~v!*jWB&?7bHa%&ma-_5fDae};hXHTWOB?}h)Ze-=QeEF~u*#834< z?edS4u$6(ep{bQIfRTk2pl4^N=L$>z-UEy*EC6?g_f|Fpy8P2506JPLYlrs`fQ_Sr zC&0+s4)*s(va$f^_;O8UKg0o`6aEje(F5p2{tzR8PV5gc z0qDg4(0c}{KlGkKfBFx-XOQ_r?-^wOL+tMv%Z__xXqsyc-Q%U8`=NU;eUMI<=Xv)@7nDD!gp;Bf8o0} z$G`Aho6}$TuFd(cI=yRi`3v8*x&DRk+T8xZ|2{-=zn>oeT#odAPSyWCR{vo|2RmzX z;8#<__qP;(f4NBOIoO%HXwbjkYz*)A_n&|Nr}bY6NdH{Qe?Fu!*@4dZ)b zF|fRMIs*d}o9BPU8vN@C_s^~Re*FK^mRKKDh)!J{e>1HPCt*@}xLl(j_!jYzXK4S%wBQJByDlKap|TaDPs=eQ*imioYBn*!WfZ&wCUQv{e~uw0SS-6p?Df28^k^T)^(7PXDn~c$>MEo=ugB9tFRPX#`3faHz7emKx%jhH z)zO>lak(x?JyXhDhOHcyEqXsy;!CeH8iK|fkV_!*~q@QdEr$Yad)A{1p zxuwcu%_#->tLZ6(x&+|;!jQT zd?OWL8*!>74EYUpy&mgZkUnOB7DXNCkJ4kwE$joc#{k{W7RC z^yrI;QQ5U5n!D4?WkO^k_+Uk*AMpxWFy}PMIDt{R<$B3{j``A74fTHHTdeNAA~;S=PVuZwJi} zY`2O1_ofm|c&WGI2N8MWg_D?yX`~7uHK|l3rN*OH1m~g1gIMP9>3F4AzaTEKp4JKP zQVgSH#JS1HI@iIlN9Mj{W`f90$`VM=}?ShZ2VAY-p&2tQ)1oW-bij|U$N?Ruuq zHM^{K#x>E9_IgP+6qpcIj=b14T0%Jc=#i1MayZDfV{toMwUkr7OM5H}bq^LaXji9F zV<7JYE$|tCuyyc9`Q%&f2{cC)c+Ct)`WHgGneOYHf>khXGX!&1+UvOde=lEf*_bos zTeKS)*#`1L*l{qo%eJms%9+6{`10$6t3+(Dky_$qkf&34VNJTa#VFy%-WK&ZJA19U zA-wCZy=Yb+>MU3>Uvu_fAQTXjE#P>0&&s)b3?UKfeD@UuQj>IukF$n?)hCsIR^9SV zAv9jYbwJXI8pWIz<3seRe|En$lW#OZd;Pb+bw%zw{G=1vNau`q4VioCu*u-)De|`-d$AZH@dtUey ztLMHvps>(V6}kcMzo){vVjbCe?qq%2aJOaEh+;ryP`lNl|B;K;1z~>6J_cw}`Dsmb z%2V2F?TEyuQ;|Rx>5~UD7PfI^Zc-efGre(OyF9HQb^>iyF?1m{#LTx@SXtV&Qr6(& zZ?DUgLfRS01;1}Xe}aSKFZ zjqT1ZappFUo3vG1FWqYEszv>^DY=9=Cq>I z&qlJAPgZZi-(XF2g@KbDjxXUG#E+~~h7Ikc5BDf}M&sb!7O5w ziDBo54q%vZZ+5j+ugIh9lI8)iv2;+N_6n^Y2_Gj$e{8U7I3(`QY@9f>je90iZQCJJ z_WFB2-kXVAc{pQo!Y{#%ikN&MP6U}CO}jN|x*3XIV@PO}+>gEagsI9qTgc#&l)#?r zl347N?yymllf3%92JNBw(GIiJ=I7!hsZhS~Z3yBiGe?TQGTLa9EBn>P#|mkaF24S8 zVr2Bie=p~KL=!5f1y^Y9?0V6%9pPt|HPicOu49~_?O=V%l|JA2niQu`RDyM#vN6uQ zeEp4_h`mGluf?+JO(6KoEKU>gnR(V=;7 zw>ZR615u5Z2%QvR!fzXzGS;L>R!Y*kD7V_zKHs=MsKR(GTam-&ffA*dyX=+fDX1ES zzc8DQ8L`B6d_>zl$3wA@u66g(rqgvsF`00kKwhCKylnp-l#*|=PMIG_h}wtu9N4vk ze~vNEPhE@L2CC1nMVU5vYq(&^uGxhN?o{PImZxb5ugZ(D1h658SN#f`VBlYz>9HH< zsGw*%Y^I!gjtKD!uV{esa|E~sKB?Ua^cot;v;czg2WMB=nzi=!41S2>8QEXhB_#f% zS?13r1(I^6(e|53=2S6^{9p{*|Xv{zg6-0gRz~ts@s#4w=`0kpfYH z8ofVG;iC+){10@F)HWPTHt^P#&A0mum5DT-TZmA9n1P;Ygr^teb)8%!z7(f&e;-0{ ztj!?JC=NY=_DO%S7byR;U;=AZ^Wb21HdL@3@YUz3f7K$0Am}@LeW?Z9o3tE zPOptR*>j6`r0@vcVliz|Bs@Smp^rAejFj~fz3MK9v~cTT*~?4jv87wS{ZbFb9Hu6J3;a;7rn_E4)%q-QPQ6Q^*ln~HY@GiXHJXMoxc>!f1r_DX1RP{FS=a@Q35py~Hyo>%iIcHIiAFsH5*Js2Z! z86efL$3pcXPi(f$4f5v+63AY1z=`rkZSc!+&D}FEU)-n_YNJiSj@}lhw2A)3Sx&&x zk-y+;3{MeU%8DnbCRJpve@VJyF?$^mp;51BBdw#t8iXw_yFZ6y(Y$|`UY?IymIxQ- z)%w8?E|EjjrgTd7{pWt;{6^$@pnQfGn%o9l)8yxPHYd(u9aMC=17>jWE$F7s%51D_ zFvhkPL7X2jzWZ+ZyZhFH`}B)5lcJL~Cr0P^Y%vB&PZ5X8n>1mee`Uuwo2L^oZ7$!U z$guTm*=^=xzhQu4jaqTj!c{R((O!c1o8C=@6WXmRnMNo0n;*nsLu&Q>Y&@8p!7G^ud4(CQ&e*yz~mL z#+CoMue_^|=vLyrUtF%48o z_M0<8^OZq zD89K#e$x$j`;3?;8p*I<aUjjA5A%q+sqoxgk0RW9{ebWDw1jq zByGSL^Vjpk4zP;!fK|#SG0`hd&KgHg>WKP~Im9EAwK7Q50&04YPBs|Y$s*o>pqGWN z@z^jfH~meWi(Q@o{(r&9AE(P0rK@Bxyr0WHTX?l+C}9LGX_Ms*AN3@nuXUC(lHDgj z{ybM7Q=WH+NWDF{Lflq6GQ?QK+xoN!J%7Kk9>QTWBfb=PCk|@X@N;B{E3Jz5g@U!K zo@cs=2KKBR@MECcx#eLyXhtXAz^XgGJK?Q*CsXUR%`TDWbbn@X{HE5x42HHBMP6o$ z>dAk=Ru4_5bLScTE7LFVz-G27gH$~;-|9_&@je|_$#{M0hrnOsHbRH8L7W@9U5Ww4 zT#28)pIN9nQu+9+@goLGJze+;oOy4oJzG0K=ItaT?#I$eTTTi&Vol*no$Dgjp$c;7 z-#EphKp;2b;(x1!l@$mm#bfOC?(wSV8v18a(^ZpFEArD1{tx?Z&sM{?ze(09d||sU2dBUf%atLMTff_ zAV3J)Of}n8?@{w8Q~F6PcR%;)DN#RJ0oOPF*x3{Dfq(lo9t;*Pfw9ad)?+2jbRp{3 zlXWN&I8Z_!e2G;}r81#wUH-7ad~>Fz-lhvQN{HEG37(d-q;|4gKTw6+XJ!51z#4B( zjeTbbaz!L?UzIT8@xynx$2~{BZ_fjJ9^Zis>$eG}6SK;r#rY!_%-m%uJ7Bt|jjgik z5gP^{Hh(gZ-#n5C*ZmqZXeL86i&PXc^3vU7m^?A8KIF_?Z*GkmZXQI^q^ZjbEzb*A zuUg@jod`bdqd6@a*|C!nWJ1baB(Bjl=9=vG{Q?l;&R*t9Dh%dETsj<@oH(|l1_HGr;&k!Hz9JV zs;Zwqm3=b$gPw{D>*6lVg~5IO+f!>|TYW;z<7xu0vMQ+!A&ia`^Q9c=EoqB+Rffi0cx)}uFkV%#OyQqiE}oqy zEPwh3g%3Khbex-IlPgs0-+?&%KI|+86W3)+Xm68F^$yDs z3ucUN)*)c;F?Ac;ZqZ}1C`T?aQWLHy0;e!3< z_#@W`!&cDg=0w&^$~uPcGN>d0J(t}EA#Z25(2*7DVi-zptk7MR-P^8V`XMQ_2Y>L( z$8gYX1{oewj!KdS)4ucJ%L#LwvA14-xmh+nN?=-SE~37@aMrWQfhL6mJCMWK)e(u2ZlE;;BvZplT(e@|eB z`os)Tvx9ypFmRKW$r7_FMG=6BEo+Yf870}Uj$W6D12}9eXw`-1b?v+Cdzis#s1*17?QW->vi1$H=<#4h8;X_pGj!&3C?wm zT_ZGVc+)*1iON5|>i$$vW5Y}HS-fROtDD)|B!n#Wvb{Ak&%Kp(6 zR3t8x?vH5GcRSEoZlJ6|_FRwrMdjZPQJ! zSWnDRWIomfMr-dkQZ<*XBE+^R%FkrgKz#}VvVD!YP;Udo5G)Ay9+^BXdq?DyDoAcSn=8 zm1A+h{z3f-o=_u${aJ#Zhc2HEI1;O|q?vRW;hw*xy{Ygb%AhW2VV(UZ00GHNw(X=O5he#dw(aEwB^Q)Vti7&|BWRE zY@;Vl!V=w>=%NGvaco~A)lXM`JgPQyxzEM%UgLz3MwcCFZ+btlo`16$(Zk()Yk!cG zf*d||c#W9}E*G=fU4|qRCqk>_dXL7o5pzcZx852@ofA~`sIS2QB zhahCd$7bQ|Pk$kM1mGK)E!ni+;0^C_ONKT?$`68c#o3}6l{|N!f7z`f&8Yv#8J1MG z1?5lYeLkb0M{_=XyU)X#Mu?kjBmIyItCv><`T99#LgP6g?NWi+=sq65jus&xLMy@K zb|}f-n`Cop9b+CS$--2J-3|QKpC<9Z@P#a8ZtvERaevP(?;$w5v^$vJFMd?wql`0H zbjF4RYElDCn&77M&5V`XcD>F|c)AZ}f-FcN%vBR7#<<>o@a($v4&(`FZE}IL8Drpv zwcp2Z>I{oGM;az&mr$GV;S{fNkr>WycSl$+)8#*cCYsj;_PmU)=1xk{4OC~#kBC!} zr$w_ubbn01>@rONYg7ssLEFFQiT52xQO3HUL<8eP4lxmJw*#=-`MNo8Hl{hic21mL znfz^F)g;ys)79SCdaw8P31MLjz{I&Z2eu@RC#GFcR&kd`a9|!Y>DTrJ7zlRS$;foC zm^FFwoeG<>Muuq^ck=YM?Bpm zNR!J#G`uYIj;14LMQ>&^=6-zGm_Dg|q!{W;)A_LJhkq=O!Puj`6n4dvs&IdD+ca!E zUqNCu+830a7xaz~kJ-15Vtda|z)Eh)dm38yNHDnS&TQmP8_qF)Dtk_xja=2@!g^2i@oYo+31N-2w9 z*Vs=E3S-2%mU+Rb|p}jS$_vixkI0UR&r1`r)!Mk5w^5vZq+8A3OM1nmw@Mr zDeP7m#f0|ic*ko+#aAP`&dkTM$^@|;ZJtRMVkNQbDzF%=Qrb`;(w%0%Mna5 zzOT`INTMX?-dwBO0pAkeWRLbV*KbzXvE2YnMJ>;BTjFoY6P0vADURaBq7f#}dh*JW z8Vn?|p_L?qXHbYFXOywL9UdVVhbR~cq!-FDkfbrmQi9?L>i#J%K8{5&Vz^vk87>cH zUk$&PwVmS_3ru}NqLqh(zJFcCBSc&-D_J4t0Ta;y)%`Y?# z%;jWgG8@FjjB>3Ze@--6Q!g|qSNtuTm`|i^;K^CXxHh%dC~(IwVn(LlDea5pNU2=l z=cAKO%UFHT!wpPBB17&^v3nO{ANS)#!De*Blv&K0D}a{1J_2j{c3&+lG;|UDszss5 z$_OICm1B`PTTEhwE`P-?EEO>5^PQKbo>G#*gsBLv$696+u|YLh&IIIEmgra%xy(3t zHtfbRw^=*z{bt-3S`Tp>IbwUiBA?=`^+5_`jbbcHkP+BS{^AF|oa14@V&tj`mmQM$e|Ys)=t)3#yHYW`XENxyGQiYq7GV!#qwZrfkv8$_3q^zbs+i zn^S)||3ramh~{7HlHO!3V&=q?9)&pSc{GG^X4q`3vr?+V5q^|1N`3k2jsBqKWq`7f zcG?_%_DVfRaerN_=y$cZ`@$I~Dd6ADH0U5ZcM^amNnvzHw?w=*A3;hD8tHg1WwnLs zFihr2ilQ~&M*%t59%~8TCyyS_unM`eu<=>HK&1ypHrW~1O;1K~ekY?MGQtsv@HNg6 z9@rhi^>HpHDMaPlju~3TLmjgd8tf)+6I|jKfKbha3V*tlTffpL$_asF-|tBtzh*sm zLhS;o-yYP%q zg}LiU@N_MYt0qtI&h2FB+tr=DR>5hkXzY9^71-$-vZ@|?KCtwM)V z0>X-awV|=-)~R7yRFGG#1l`tmhP=vV*0IJ8^B9(uG^QO09Mipu=sqa-KfuIqhN`QS z7T8rYP3j0=v=|EUG)&(1gm7c4=nZq?HXpf_*d;mv#+J+_3=_up{)! zEw>An?<`V0Mh6Esej3#V#4!1r>_c*k7_=h(;j9x4f@kzHn|h(+Uc(ZCSV1qp3StMj zz<+-KUWrOvDg)sbm$1_2cc|GyLs8o3JSuDcqgS8kII~f>Emf#5aL9q`InxdhcEh#P zQ33b_^!!zv$XfTk1Zam!^&A?(uiD6{e)k7ttv0$9KWa!1GFx}Gt>$LeIL2%!Jv20* zANfP$yhR*s$$UaOu?6EmvQm9A1q$S{lYfHH7xjB(z%_o39j@&%a~zg7;nZMDNep^W zBYA3Khls5X=po#zl*~;JHp`z#RdA@$-A>6T2=y2q9NM{JtlZvBts=uowek6kVdEd&amdyF8S-TGrnFvMnGKKCbwZ2qwVJE`^)(A2w z*J%_@477#}xQBOag@c|i-jbz0dOM|x$swUzxG0IahcsAzm9LOk=@aryuvR;Ca9Ke2 z7EGpn&W;HRYzfsMtF57>gZ4C=N`IX2t<%XpPR~s5$iwZJCp{;ywDdV9BeSMwdN1%t z(xdaQln9el%E9oY6@*vdld?BUv;Z^8Yh2^pYhp*fq&0<+8 zqs|p#>8F)c)IC~oq#jvHD5d-tn*5kMf%Ok+qv1u^9KrHGx3OO${9s3326nNoZu;h& zEJLT4;Og26bZ(5g>Y3Fu7j!;yrK@I1hrrn0#vRVBJZ=iFi!I?O;5h3j)&?Hti4^XP zpx1-#Ec^A`uyt0H|LPf}5`W_=g_$ztH!9ZE(ShG=!2fbMD{_VzK3ejOj}4>1w1i(8 z_2GP`KV11qh#nCRC0>@j;An03VA?_ERo;m+X;+oVpGkVAO{Bvmd6qzzmeKrakutPJ z0zye6QC66mf4zB3Fbbo(jlMdg+d(Q|h$BFQjwlot5~4wgr3Oml$A6NJA#-D!5;|xw z$A^1=C$BKAj2ec7K-ZqN79Mu0&hoE|1`OTm$jyfw-MdVpk>eJ}8TIHkG7{dyNgo1= zSk0V|>9RIQNq|LS;4chTr7|bWxM7c9w&=p43VWkgrBNA>*9yaB*B|oUO6#q(YLmc; zA{{EJsI$x^Q|m_chJTL9im!!*V>&jm`Oa^p1Hccm#2cGe<>CFvcuyy}c+3tlDlK+; z>NgI8<=RpAoJ@};45)mkm{Y{l!6VAHuiD2=Re|N9=C>9finTHrLxjwQk%Cghh6+Qa zDZ07IgRwq7VwQ;bqFXuDe=Y!Nhd?ycI8m!wMk_cQHM<Nchki(um|d*uld6G|91NntST zoz%hBR8KDwjXko+$4Axo_mPj3;^H4Xj&pM(_nH9)1#%;UUQJWo%Uzc=n>fp9SayrN z=5Jd;Jw}^X!GGmA>khtZB)HiVOC7Afi)Iq7fQDWg$v$+Mg0=YQuc38kI2S8x?6#oym1NLIL( zQ;~%nu7W7DmU2jcj#2I0i)00H;r?w@Os0nky#3RR#g;FJDdoHXhD?Q!oUg~S$ zl5gbLrorfB9$`nk_50q=`*Q-^?NCmSsKE>C2%}F+#@C}t(Yn(eXZIa*sFmz6?{@Ga!=g~1P&NUfuBZGUv$ueva&%tEe(dYW`BoJvS%dby{E z%h2}B7!G*OJ+K@_jlO%r5@MzhvBBb5@y0kJc|UuRZ8VUPet~gqU?5iqnU>}|G+<-mnx6zwSpVs%Mx4l2HL2ahgNyI?ir}`r4!tW%Nsn zihnZQaYubVX!&+Izq#n;7KFpsoaQ^}0_Hr5hsj!d2IS8M zZJ=7DP9!yF>OrzrKp5GDfe);P+9e=-dh0S}!4NJ?4-;R~G5P#2^ht|MTHC2*$$zgi zLD>wDQ<@FBGB?Q}mE!zwb;%~=d50Wf#Ze;Vr7byc+S5OCV6d@Wvn`{S-5$jdl`e*r6 zx4j2klR*+K`yW+qa9^=a_qgs$k$-D=wVN6uldTI(#GUN~E0->)UzRr%rc>N?`hGd! zD2x2u41v&_GU{9SGV6ctCPrHeje`&4-M|}xqak>Lp0X0TqWx-|vx0EtILKJZprp7| z0M#VgY6s2=3f~y{#W%k%9VwP%39SeTBc39ewwIEUXikvBXUEb-ab@fT zVN3hl0WZpQWP(k(iq6BZhkyFa_4nNjY=Ti=1#tSf`ADS|ui4Kr-)7 z9?Sdea75fsJ4u?i83Q*DB2xmfkJ{5u1+B=AtAdMvkrlsgWH5lKXn!`!4{bZ7FyN0S zr34V!380g_O~rFl+;ZW@e`DP4-8;{63klCwVu5cQiHw$UWD$aWZx1n~B|LdZ;K(ey zEeX(sA&Bk*ZzRYY)GPj?-zem+Q{|CW4{}%2=D-7ms}n^@q!8}&3}Y0Wm37v1p{!~UJdM(@?BxE@5(%gi^!`rpWG!SffN(8kK;p< zpAXaeD7whk3uFB5Q|Xkeic8?o1{jAJE1RY5u*^EpvZ=zFjkj&0uktM+pE_&HnZP@o z@lLx=y3UGbgo%};K_8^NjP)AZcUrnSXAY#y+@X#y9~R8yZ3I&?yzX z>LB*1X^EmM)#A6fCKZX=F3qRXX;o`(TOpcudVxHzcFF9wYCBy69}A{LS+I|#Q@URP zbz)fQ(CVH$f?@n9p7xvwJRzU}rSEKxn#C5xKj+hWTQXi|D-mam+;vvWx509K6kR}q zu4JB;bAKpquLUu{c#mNPgs%C%!aQzKNeH93Pjr(A0ODf{yfsD})$R-YcLy;w2N9r9 zZ2||7Kx&IAdD&l;((Sx%rp=JA_BB}#hx5vmu&{ZIT#VEp;397$IKnl|%xJef9|?H% z=)bG+V{9BJVhy<#Ru5_2$rtLWXy5)iC)!+#9e;QYyO@H4`#c+y7~A=!i1Rtgg1OUQ zHD~C8rM_{eUcniDWe|jgVURN zXB#3$3M|VYEDOG%7-|@q*iz*_tmg#DyoIi61xmImX=lMjd-yolB^WCLm7k9B%-nb+ zt$%3BXZNaPewWBI7dKGwLTuXc(9`}V3m)7K9FD>t{~uouAi?`Z5Jgi^k15K*pM8qG zPG_OqX&eIBUqbmpim8DH{bjYL{&QrNnS&y9d-M=uo?rirr>}~NMZKw=gNKe za{VBGza^XU>>`zsj>hed4U(Ymj^9#GT8EB(4=N6D7G=jQ*%(E9B=avZ@ibNHRevZk zI7KYxC?>>nqUoo7qKAFBou#z#KP(32_HVol6~T(HQ{yQnDnrBOGS)B$%xrUx(ksTA zU$Y>vPDj#)M^B(#7^M5QLrT5&Ftxs@!*G=dH3z+RjCO3h9bZ)VfnOKhnBB)Em1!Iu z(y>USQ_8{c+D__Jr$W!IsnS=^hJW8${5qZcqLEqh)p*KR08QGtx%w1VNK_S`)3^vJ zzg3ix6S~xG_K?u4B8EC>#A2>x-OuO(XKDd$qj`~({h04#L}`aM6@{;`#U5RQ^oZ_- z-Icr_Q;upKA5qs#dljP&w5!2jovVnYIwPAsoloz{eFmN+gzid(w{O>c@qcWg_!iRI zekO_IVAM$Nb6_f&Df)rtB*~df^)6wnZ$@*JB}@UiCh0!oKpN%lRl#_)=IghY{kz2q zU11VA?gq_A22g|3X-~$VoP_%PDVElJXxlgPEXsC~ zXc5jc6q<&e@uPnv?`xY*3zYd{$5~6`VJ76gC#u43sOt2sRI?}PDu2B2RPa>4RFMb0 zcZ2@GB?hcSGd$#{K=`5DNQvHBI^Sjc7zqN(dlxITf_}Cbb@`|1B6n%(s!Hr?2q3hBNvqepU@j51g|FB!lL{58LEynv_@7!igjILQIvUy z0R^a!6P#ii=EULxd{QFcIh}CJ3VLdr2Ai^uC@!_)RE7Kd;eS3@{!%fm7Nkcy@m@$T z#gA9T;dW}e3SPjj1W_cehMh~D6XLln8Q-l_D;5}Oj#$gQh8p~_`uDvqUycDckCbW%)aVAsP<<~W<7NRki z_G!gGD&30(4nIGQj^dd&D9&mhO+iw~vRL4S8P$Rhv*z*JuM|!};#J1r|JT>QEZ|Cv z451vYQUc=oSAaiy^f~YJ661}25XOmlPzcQ^VI3YW^?!(_Gr_ExqPpouY}t6^r)yAqMR#pn5*dF^|f@jSHtN z5#R6m(|=x2)b^zAs=N9RquzBTuT@{&UgIC+%e$hM=8(9b2hms;tYpqD~Io%j$h%$ShX}hjc8#Xjq z#}M}L_As{wN!r2i5owYyAWULcgpaDDUD@j}wcVEcKz zm=du#=9Q&Rf~JC?jA4+zoslE0mFZ^n(R~QIMY&gQ%TCK}qLH68F3-Std@m_1k)AG9 zqJQGFk;muECfe&(j^{9teSQK;^amWV+-^Px@FhBq6>C1eG^M{$Foz{y$8Ub)cPusE ztPuwCzg9gSxiGuxD-)r>K_ZBfhwds4g?!XaSY>8m+?P}v9Acn~KWuX)naQUJeuabX zLm>NQS|$MnSKUdSl9qjhrbTLQ_FAwSigTjaJpcXD6R?NA%Vn!(mwJL3PEgS51LpW#k!S>+*5TgghX3vIpq2@5+Y2gd&M{Pzl&d~*mV^+j-pkk<0`+uv( z44wuWfMe(JPUz9dw<_ud&*rdD;g#VhqT)FcaXM?ZBw=yHlZ=GgNvRVQGvTEb8RNco zKRKS4jIn@JBd&=IZDfdCIN92l25Wrpnrg!RmE~0j+w0NQ_8i3u_qqX@6%lV^=x%Qd zf$OqONqoso*TMAdZqIO^svHeme1C*&{->6za0@me;Sa#iDABF&9u!B+*D}MN?18jHQa5v#YA`DAsGMY*vQJ^xKY{!-IxZ(IiK6@ z{DVEYAf63JIJXi<1R*m~-1*w0cDCt3AkgEEuVyX?m1J1s72UZnpKdHM{F5$ki2@c{ zEj0RE%*#!QzzI7a)^Nn#)_m+~BKq{TvJTE@eW+C>Nxo4xt%~_(xPK5nTyWnbR*i&9 zw~cQC$G1n~P{HGyNZlV`pnEjN6t{GOx4_$PfFB*FYyM~KDy|ajwqzNb%atMuRGZ^` zHRF~Vw>ORzTbu0lE(f;*Fy<{K$0d?X4@?x4UnbNfyh zzA{oKkM8046!VDTLVs3pmsWd%;L|;-KVn>bR0eUltr!(8%&mg|N*TBNXU+#rf5XzM5z$mt~@s?P^d51YABW zwR(@xdK}pk5VqyL$e(WDbL!`CM;9z#JQCwgg6r}spbb^Te}ANl_N%kqQkY~b#!H|K z5J%Yi+w_LLMIC~inS+I}hO$8!0t>xbKXI$n{LthIgDy|B`S6QX(0TVxn^}ij_h+A6 zyX6BT&7ESt_0E`)t{f*NS6p5c=T4Vsh~F{{-1=Mn0OGvjkw@O9q5zb=?hhL7@&lMr z$N7k_{w0A;xqlsO+sO}2TMkbo`TGk6v?Zp-5exRR=j%RIM13*4X0QSc&6pWh-J**Vkhaw6Nx@PKpckrPC!m zA!ESHvUUD!fg;$AAw023I0T!_I$O)=*ZI?Ydzv!;M}J2HxB$i)4w3K7$6-IP4W8}X z&aZT+`MD?M+kCS=!NZO6=FT)|eZ46PSl0K=t;r=(29 z)u2@LjDO%xHDNDaMb6G4dkZ<#3b!zQS2iK>n}s5+Oi7!ig-krB;HKxni$3P7y?bOI zEJ+5xX(6QDI>lZd#_M%PxHOWJIB>*?;lef-sbj02O;zQtSIxe-!&u6(KzVFg3+Ws} z727Ff$HteUj3R!(MpL9x@@q1i31DM;`yt67!GGN~DYL1>$At@3wJ$VKru1X_nrgig zv`um1JYksadG@&~)_lv@tbKDwJG8_@Y4B9c_x=LvC2{E82Vbv`JV{4`#VcJr(d)BZ ztA7@dN|E!9dynCE7D#V7tsh}i21+RKC#)9VY<;$zMbnjCG7!-m%;oIsXBjBfzrN8q zaQj4eoABj+m~*QI=;F^+zb51CNFtdTRtEkYL0-{6#z{ynlBj@xl~%Y$HPxq69xhkC zux~MEkgT=NHaYXl(J(J$+yGyDF&E3A5`QaiI=?|@x?Tm=Q>-+8i@?(`@H)kAy~zZ=;Tw99itHZPEQmo;VAntR7Rg)k_yX*mVFJxBrecsy8w%M=9g#qnm4=#V`avZy7gT+6@TQ# ziZV1@=9n2iFEaV$w$+l0y%Ufq*6)!tqKuE$+Bi}M@qc&QDF+ax#-dsuLSK)aE0&(e^x1e$Kuxn&NIOGRlv zDHty=M5W{Yb`2_1rrE)(U3*!ha9i~3-Acg+CDe!7=*n@vxX#c^sow!(?tg3kYX|^T7tyHM3J-ynes)lI_IL+h(;Lx?n7z< z4Akc)8RAF&<&XQqEp#214mgExjpU*eO5PL|6*)ZNUyy)p(Y^AvPnw-w*;5({YYI6K=0O^cM@0T&fYXBA*4-IoM(`S_QSHe|sHtW96?i@T(?w)O)w4Yw?ZQ-h zjStW3aBVh0vR&QOcYpZ`=fF&b#sDAI_+b%mBXGfMX}T}fS0SiqMq3cCE3*g_{TlZ$ z9t?R4nRU@==#94G^?IYSV#ugfHKMDEpcW%5^6ek}l!*{lm(>g54ctuCJGb>#Z|KTk zn|yU_4qq@LKm!TJ<$qnl4%~kfd)cTZKZbajHa}bE5S$|5&VMtr*jZ(LGq4i*If+zr zG{anR!SZ}ut2orfTG>T2+E8QhD{voT*yVZ(FuFjyZOQ}fAMM@{q`}??)1L-k6+pQD@q?NccC|TG#|VmQa9UtVzVkmO2spqJ}-wKBx+9=&Jd=>42kQbb^^tX z!m^)$0?H2kmwzbR`p+?f1Ga*Kl4quvbP=|4BtH}K`}>%n1o%^Kg3zQ?Qm`BD%POs8VEJOmF!7L$!#c9RUKjf%}Lm!a*-hjB-S@ zA#9FJ#hsOA|Cc#O= zs_yY>&!01oog?ifmx2E0C2_~i1DUv)te_IXJfT;2%?wcO1*g%&CIo3_jbTPMsmZ1T zrQ6=;*g|{1C9fg#O43Ff$fb9N6{F-D;`DWC^Pm zaGw`9i+^b{Ok?laF5Vt+$XFPoV70A_OPZC7R1$o#htgyxUn*JL>xg-|k@P=aphD4f zl$dtq!6d7Sm&=S}h;=&Xo=m{Ew92<^-w4?SaSFa9#Sc*Vr?6sv`j|Ph(%3xSmo!$@ z;autmlXD#h$o8+AQc8|U`?T&JQB?Z|mbq^u`F}0w&?kWy^O7ZI+3Rt}$e8>Ee%;-L zLfqboaXbG1YvY`jEKvYtTej^k+qP}nwr$(CZQI&qyUVt^%sX)-&OFUGWhhzfNr8hPU8hV!d-Hq>q7m~K@+da4h5IKPVB_xo^iO=i7h&PQWL zZGQu;J_=4hI7{TKJUi^mrd=haSF%0!%e0PP^%DeEaVT_{OgWmXQ?5^KKk2V9_mpz# zTmW85CHe5Lhpxl<#P4b!?)W`jVuvk=!$pU`ZSCk@?D>C+mPbe{y~XTc@tAG+yfSAc zo=h=Fd=^|te1XQS|82ydvgxHYMIKpM2!Dn{Lpn63x6FS1->MLmS-IW;*<$vwsW_>WKja=pJP=71N z+^0FC;230t?tv$3?(N#&7h~wwVVGR?IbBE-+&3yx4g*Ns@IF6B%xY$#Qr?Y} zG#IjjHCHJ7_l=fY{+yR{&htk_WgA9mfvrLz9B7VO16`?|Dg_No4@{)V(T<$?M&(T! zsLjJa8$A|*JWw)$lpp%+;@ zh=q3UVDcc?DQM#vT4zC5ln0XS(+5XzZ_x`Nygprhr<=AM~-psfj(Q_v4SO&ZZMi_%A2!4>=QPJh7Mfg-%S z0+~G-2t9c?pU|Mz4&N3uDz}SElwRT!_~+1 z-R+K8AqJeo^gs?h+>S8!tlK@s_*YHw2IrZGI%a_>o|M$jL-cBpynjvnwIM%$yXb}d zpr|g_`3($y+*_(<##BYWqEph^>gg~^2t0g`*J^|8eA6dPftreB%85ia_mHaOG=Su; zd>25CyNLxEl8NG-q-j<$O2RnO6F0XaiZW}WYkqxqhkn3&nyk$p2y>``u95?1A;H0< z1Aq7B1@kX^af9U*kbk^~QODLY?RnU}spsuctPBTT`p3E|yWqwAC3-?SpJd9uYsO3i zH_r1K<-A%-zDr}M0-vD(pf>6iuRdoeT8*uFZ88b~YGjJ*-EPXUe*nKjYk2 zv0CLQjwIS+OK2~PKg-Cm;?ZxN#hCML+oyge1-~DJ!Uc<%x_`p5BGzZ64~0{qJ>-sY z=bs%dG8!?=7_6THRZ31(~Fac8l9?-}@j{3#_ybUFj4rEjSus!y~R zO6@$(*bhsKc^QII;t`*?!}Z37{>%5^{n?sv1$%Hr-NgU}>Yga;ktOngjerVwE4zX) zFO9a72SEzQ9e)glZl(9&Dzs9C(VV_ZCWng1kqfI2%{HgIvvRO5oSPS#=!FD!jo+1DT+sDry# z74}P38mBo^g(n9jwIZx#hq_mG_o$$0EIDKvR^a`jI)6~pB&Opu(Bm(s7+I}^PpQz0 zj6!n@1&zimN}_*jSke=E57}Jxrec_-^{Ec(?50yDq1T^w2u`d19J5LBT8sxaitc1_ z@t39@LO&}iz-X|mZJthO@Td}$qb4S?JB@JeE6J9}wENh>lW6rT@;FW>FvL}XY4T}* zW_NW9M}L7I00ihvqNi%8jGt0+LjK5C`Y^?%KP8aTyp1S2^D*}RtJ5=+rq+I@3cb!F z&U7a3--yV7T=Y~R<+U9cY1^1s$S;G&3MCo@(h^s?v{u;R`pq9gzc@p$ydP1z8)YFg z)j6hJcMTTALjwQRE4F6L@7X<}beukmodybb?4t(OK zP=8gRVOyH)h5J`4ZPjsna+SQ~8HIp!oZ^pzOIIIGzgackk*C@O)Uf zoA-EHQg&~=XLmq2611A825qTn8uy3Iqd$fl`?kZd;Gv;Cv<4~qnV1dlk*vZHs??@RioTkZ zmBwgFM@c~E)>V!o_ndxq%3V2sKrj*M#Z=ILj_0`jhfp&<8N)3e+&CC)`p&@_HGf@{ z3vFE7O+pFX=sAG0e|dw`Dy6{B(E;=_(0xab9qDV=;y;G6d$o9S$97-SeF;VLcu!xRcnm2LMj2+^(`)CW;8crkcA{*KZx(EY%f zQ!mz>nO6v=>J&pV_(mzkt9UOOUw`XZ_oW&HiZl&1Lpt{u81Zxz0f3iMiG)ay5OR)_ zO|F(xoyW(_9@NuUToUb-V4J)Va4XVl8$EW3=}%SHgs7}Q^PK}RnYA61QPyAXeS?lxO;io>cCN%A$ zwh%^^JOm{-iF&rx`}*kJ(_VrrGHFyc$Yz!Ls!ZXy7E(RnVR5ywXjxCN1rj0aU?)~! zxDWySeht!^wX~Dvx>3tmx)-(&`)BseS}U-tQy}K2*EV6twQ{!8;eVdEev)4KE;JS} zY!#j$Wr37!h?&bEQh+e)J%WCFsRU@O+<^lh4uV6`hgh=C=~q5@tJRYG)8#wv5K;Ce z@}gc0iPW#LxcD6RR^{(2gwxU3hsh+r;gq#BA|bX%AJlW#Xfm?(AuMFNpkm0UV)ZE%U)yk9qZivI%s6md5a z?Plpy$DBqHgVMqaMpZBJ`^2xhhclU{`Nn#yRKjCB+}?;wx_|G79GePw(i$__X?Ts(z{u$zwJTVCxo1C>i}#q z_IKX_Y(TWsn14ONGij{aXduB8?jaC}j-uU_AX#AfXGeK9uN3ww4I_z(;)<}wgBzic zxRn8HLa{lQzbQ)V3Bw6Tn^sMV1%?X0Cb4;K?`JuyH-E;uclv*q-LaM+YxAfd0JNIQ zlkNo;kHTFHPS|{evdCU%Cn$(X_~akhbZB}#-f&;n=hc>CouFoTyEMX%IM5o5r-(jF z8jXG`L2Asl`sYU^QCc(b?D2W1!3wu5Vgw-@K*UCBy?Hq7&jM2XOX-9J+zq-yS4W@u z<57fLQ-2iK|18GP0T0ocL6-F90?aBy1lSctBZ@;TJkWUz3fA3+rfjX4;jQ+iRWoGt-2su5*5v)=_UC^N9``#G_NU4 z7oJTzD%gLm`#C(~@E!tf@b`n3CcDxG2#!MO-+xvLK{P2BN9@-1GQX<#0)4}B*XrH1ObI9asPIwdXu!}!p9ayO z%@7=c;Wi-ymUOO#g=&e6$eV3DQEJG)__B3{s$PATA09r9jlcTGt|f>VA2i<=Og1W} z5q~%6ahzcBP_UB!Y7-Oklo!U`6*PY|2f%iqrEq~rImcBMcmA|=1VFy z#L1$bk?a-mn86)PW0YMFo-g@AGUfVb7Jno}w>t`x+duv4mC_tQGTJo=rFf+XRaFM5 z#}K>z(b(b;^lnRxrRXApxBNhwktj?V8>Y&MQj%!N`Fi{ zg~o!JoJhd(!68R|784$W?JNfDa8-X}U+ebm#)A&;{!x0x>XNtvllEqGg*sA0{3*cy zBHGsKS9=agPIp)`Teme$SYz|mx2eZW6YAL^sdF)5S1JS-cvA}KI;#e?Rm$}0BXoHY zb7Srb&f_zl9{!$vmnVyi>D$VMKYt6$>84{-?;2{^CvCOx4zxlEMhl$A)~?3P%!1Me zw;wsZ%wEsKzRdq#@kDQ!vI%H?ZTiKGkgk|X25)|iiIpa#Rv#}9OVJfOQ9#Gqd6;Yr zhRZ}iYq(gD1u_naa1Wb@G#IPpoK#yLS`WR=Mn4--Teu=UQyMyT%svDLFn{1q`$!aD zsTn`Ps~0On4;yb;#S|{`b?z@SJq&h&U(5lxiUi4UMfB$S+7|2j+2ul}#tP0nX^)AB zT)4UDcHuydu$Gau@*Lhw>C`X~Nif2QiH9a+>X3wcyrjqk5wC6RlIJJ@se2i9z2Umvh*E#gDqJL* zZ^ASWnb~0Pw!x9f2y7*ta=|@Db%FE5r`TYV)C#kfZ)2P}fYzP1*9?^4l-rMFeKmcL zGX@Zh7CEN~9ZA>^)-3yo`W%(PPN3ZRQ(DY@;7N0)GCA0(_lUR*=>=&I=oqIj2UWuD zYGNwLj6OC9Pv;A_ldMs((kW9`O*)BzWyX;*1wvRGtem1dS+opQ10g{*3%sA^;2FblQ%(+l19X8&`aeR3S~{B zub1V#9BWyoc$3+ZNsV1e?mM1Fp$sd%e@rgWthPM{+U7d6`Ru5$E*gJ%<0tbSjy_Rq z9J*d?uHSWx4RX$LBob@`#)>LSk;^hz^lhe%Bcmr&ayX>z61uSOu3#~Sn}rq3MFWB3 zs`H7rz&Z7?>+tn;4J16YEylA0h3t1D$c>rN@RRYEaOjeTCnDIhWSkite|DX_=8IcY znZfs-OR&sc^qZ=f(s+LxZ~oyTiWz3=aM*c)zcSLYzXup7GBDtA4uW|t@Y+28YPZ96 z2u_aXZ|)wO8m~bCOyenxQ?5Sm(4l`?2snp;JRZ8gVIx^IdlLr0BIG1d0+`ABmWBH< z%ujiN?PELXu)+h6@ov@V^5L!H2cv!ihM(FiDDjMCh*N}!*F}FibsKGXjbN{(5j9)Q z=>?0}9VII7H1t15-apd_dL`s@uDbdxy!V6pW8cgKLAljTSXre~y-APrQ){c%asr4= z;CXn`IuaizN<}F&=x^!Eh~Im%VT(fQ{oAl~rhah1p|aL--y(R4K>{Dy+9!WJ6B*Y&<)Y^5kM%{DK228BA3LKxxai82e+GL0k^;QLJe>gwU-~)t z433q4oPUHBhX2XDgS^)3VC@A=27%a@`ramdB7BJFlYqm&)qn7f&fDdh-vD>2vCbe% z$4S*>rv_T9I<%5tFTB^_dub9s78;-#=jY`0d#vf^;=6x@WC@RBiHh5LAKLR2(ScyR z5l`1InR#QQ9t{@%8Mi=nyM#MLLOQCRko3*IF}keDznabELLf9^^S38Ofg3w!CU%f( zZ~MVI+lAYST$xA(A#OEVK%AIi{J4_+tRu$9;!h-tD&_uJ+u!Dtw+Gj&1iHFbh1e4} z$8wiONcn$MiKUR|HMOjaX=Fbzll1;Duj#)LCMJHKmW|t|65DQs`?h^gCS@5t-1Q#l zHs;Yn;x|D?%6(%D$5em0v#tn~k)qpjrKj*<@f`y8gu&mKgX+!wmKC>@_^IEo?=_oD zJf$IvCqK$eiUw^7C~Y?%lix0@N1dad)#e6!qn>|ok}_UY-f^dcuweY3oH~(w7JbQb zoCVTsbkO{L;jL@$LL^-65dYkXrP5xq`jPP?x$sig&2&(<#&p<$d0r2+a9zychJh#d z2KhV6Ch;)ruufQutTcaoPpa$l`V9>6(OacEf}UCpX~ zFvZi>p^ZdC(prxwoL8=;QEZ|f>&+ z$$^(z!|2ukG2t%Z#!^TS7*-~J_XXK%W(mC!M8@=N?><}cAMLE|0;F}L3Rg@bj-}J| zeHP(D0=NHmmt{yE==FfSSBP)!Zy6jcKZFD2y5E;iwrOH>e^-YoeTy_1kq3;%!cxxKN8 zsJdtHmy9$nqVh;*tRrJiR5!|#5SKk z7a~eNym2dP27mXNN5-(%8@Ycl==z=&GG5+`hL#k;^J}5YC4ID&nHcmoJMUYA8GklY zzJEnkD;@A-M9nBDxjmaXA9p*aD@*Z?0*@|q@R%7LFv>;! z;QV)f%RlBSv-t}07886z;hMQ5!>CIaW| zsuZMbN_GWKR(2Qk@okmWu&bn5KM)b~?e!#2i|$nNR<#cH{ndwhFs}Jjr6ravf7vru zP&#^2TnupQUD)+43;h&S=eZU;T|4jknC1S=-4 zB)TZmT%WpKMhA9|^KtC%HTX6Vfv-3uPsQhzu_ccS=|GY{uth-#%GqSc&NhbU%L&Wkkl?0o0bOjQ2dPwL$+7K~BF>L>iC9lha9JSc3$ z_Ydu%S=tQ$c;|mpJHga<_ep+AM`JsS%MIeMoX|D7t!6mgC_>uy#B>4Gry>paNxXXzY9}X8q8u)H>gC6qxHp{^CW6Ng!*h3#;a0EE7Gz#&h0bUQPe?x2TIH@0qfww)($Gm3=5U278_}xHll^r_Fzai}Tu>zpM`NZmD?4QHA~# z^Osm`?~fb78IjvHYN-0xm_-?q086KAI@|#5M7`at05;SHftj2N1m2XJI*@i00kWl& zj5zR%n|q&l2ct@~C1~`%(^U~H=m$4|x(DSaAfx;T2IPW+w-_Fxu`S~-VH^cN`a^OdP^tsAM$?FXAMD2;M)iP zT0A8EkC1)RL~N4FJc)cH4+TSgM{jbWSJ{l3Pk4Wd`v~~)qJ!B?!Kg&Hgfj>-(k+&# zQ$PK$tu#Ya=|{Tft>o{zZyKA4Yn#F}fyB%QmyLm6gOCWW7y9~@ZaHeM+Q*oYHpn0U z`qzb$?BGSs9Stv&yH^(MYQX8Nc7^ZU@U*Q zxW(VjJPa%z>D)sZx&YAFK*L;gEmDWv1CsTSdhdgdZPhl|1{i+*vXa2rMHQi&YWEPmd+VleE68%f?_g zxir!W`2vIyi@9H1P!`gomdfsZa<$vYZ66izYeA7{!L0*y`eQV&?AX;Am!ehXR>JVN z_TXQl-_P*aYqw0*PtgGq5-KGT$GWw3{W+$cKT8}`Bw;SbBHa0m2Wfv-CV zVP9n)cgzeyy`BY8RdRjf{`#~gs;B0u@NsKSi#ZT5dj)x-=U1yNuSp@^27g%u3a%Jz zsTb3r%Muv&`}rAZoo}Kb-A8{V>z+}?t5KIu34|hWvfLpz^&)k$^K;}5iV`ngWc4@_ zoy`VZFl@TTWZCdUD1J3?+x^bCB#EqApe= z(Y*6OGV8rdeG$>`GC_nhsFo~B_C*S7NL+2m+~bILLhAI{1XQTsXS;uc^;pVKnYdD! zH66K?ER+91r|*``pFHUit^0J-+(|JoQIS%nKre4^NJIKx_vgG6Za^@LsbRwIh%0BA z&cbub9_duveG&Tez`OWdoF{9E5Td^QknF^<3rF-oGfeMz_OhZXZ-9m}_lgZQ`z}Mw z8+qqJzHxCcTo!7y;>myDvzNyv)pjnxWwyYom3JQB>RC^{pA(KP#!Hs|6qWXVFL+@? z4gof+Bg(}rb^Qh4PnuykKzZ&K04K#pT-t0nrp;81&9y-30x~{?6aWV@Dqo%s9O&tg zY_yKxf7(J$Y4{RI>AYpE#$q>j1CP|rUesl#O=o+>9|FZVr#^p`YcvJox$>@8s=9C$ zA}1wRDfbAS2`p?_#%%R@6z=?`H6SQb7~1lUy7qe--Za0U{-y@-utaCD%}PK}4+xVw z*=}UVm#|x8Vph>87H0;-Gw)C(Srn;US(i)D(#WzwnVv`L%^VWUVEF3kek;j99X(c4 z?yvq~U(RM{7J`3Ld)DGZBQ5rE4PLr4#Bez|H_-p4kJm|r4NKZ*vF8|&lgqstVbb5u zYq_#iTw&~!oQXz`f-lg`RScI!QezvQ=}+iSVIaDtxTCQpkZ2?_A@2@dmKfa?HyGnr ze8{Tc#g!koW0KC6{ytgYy}Ef~iGTwq8JVc~*PypA>IQ%NwMJXbbQRvL;~-()V`v}2 zf{rh+1^_z0Ln4lkqB;3C2n%5$hQp!r z<{niuL!?&Z2n2ZiIxErpI_KZ@ftlCY^#qQhPL&E7YD|&50&wZ{#3BuBKae^4>jBm>pJeQ6TN!K~wVWH+Np*z5L!~m*WZ^z2Y#TvuD zA;^8o@=6)vIJvyau|2AyhI!YmFYVJxWifhjSc9d7@1@%lA=HdX6A1ZFIU$=}1k$hr zH;K=4UhW$LwXs7tLlfcPNzXXHfKB zuH`~^kgp(b8(H7>-34Gb%;44W(R{N}v zeWriht(5e}0uyYfPPLA2*9|AS>IaMEnp^nj5{ocU&NI>tJIX2K50~*_zyeMAEnS&1 z>bIZ?%@%BO!jE!~?1)UVks@P_arwPLP3REOfVTh#SL2I{Ao^X9i*;2MK4H#bw$EpL zXcbNb6TX|Q@l^*!!i$e?oY=#7k`@}aMeu(}9(%Icb`MMQ{bPw`l2>AF57!V2cvhiy@!Ap zYIoKJ;HsE(EMvrzdzEf$nvGGDmNc+R>#DxP8^36;#{Q6|gqme?Rj6=GY--t{&_#dY zqP{;`!qfSB1sR>wLebF8H_Nm5t?#WRRZ=!g!6BV6o}v+%2d#uLXP0O;)$*Gv_@|euMGp1{7y|zxaYejwRMD4=3!$p8N~g;dFUJAgg!&5roL277X<&F98&Ar?uWfd`TO#xwPZY z#H#xT$Q${%hK+sura=hTU#A6yH;`X^SWZ&FR?#fN%wn$ntTV4#lfM|1=P8Dj+b&d; z+0F^7?iD75+4iTOLS*;5!L)yD;A`-290g2JloH0Ua&|;FFp8Lb$*iakQxzz>D8njs zVBnOPfGlb;a0P_co&@87XYkQ(FDo+y2~EB~wZD36V87N{V2GMNw`&n zP|~I9A;q6-V=h9&6uPD|b~EM<@Uf$*1ZC-T58%MpKOg9`Gw*OwBUw>MXIT;25Ovs$goJKj)2i?KurNFNVN?Rq*2 zMxRl%_zh+c-08ygV1V)2uu%M!Nyib@jp^az%k>}m=A4miXfX>Xjj}xzU%IBVi_QA% zQZwqdq}Tjl!^y;GS~0W;rgCVwEdM|mi_T8+>`F!@zaLbtt-&CHz^H!=o8&xpPHnt9+_oL%^r&ZnpwWED(oqOB1r36g&rs{UZe$3i7#7YVn=kd zvWF`s6d0e1o6CPdG4E{2f18KHQ*I|2bYRNwG5H_Z5PO#Fu1Ac^m80PY@1d?t?eir4 zJs3QyLBl%?&gmmB3qy)bm9Jg}xweqOC8uy!?SiJ&ozzP-I$D5Qa9l~?u;4~x8v`W3 z&a|l1Yk_JL$R+s)fyUEc?t8=nh1C!?T@u~Zc)0<|QO$ptXcdM?+$$(tDH%f5ESr~@ zqe`O&IHs8Q&7A~WuUfo^l&(^(VO(pOmb_BMeV@FmzXlH?Sf32r6S8lA`zhcWV;gZNGqT2BoA|7eoK682kY3rM9Afx>c5Yj!zYwo(-=o5QZZyeF?oc~$TjO* z=JXT(8`gjAv?)i_zH6wIj*K8@i63FfIrrS!yK?M<_>xJg46bX*e!k^Mn?TH=4vLM& zcB`gY{qRN!!MpM_l^n3rZEdsJgku^1(ZO*LyHY7Jic^k)Nr)nZlAP?#8Yq|Iwa>NK z{e|J?5@Jwgqm>!;al6o6P6zd)fH~JyPjxFn27;5E~-3A8UWeQn+>VNS?!ya$?3$qsxVHa z|IhQp*_bOxms0CEeK{RII@_;zSI&}3BL)$*S!Vn}?Vm>=uNZ#cKBY??4>hvM;&(sf zXu5w}rg$NbOKGqAGdpSd(Dqhn)+#GY`t-okO4QSevm z9D2@BJCjkn zB!<(u>~1&d^e((pVrv!;QzB#_iEcv5G!K7M-veplVJ@W8zTIa7?Gj`9!*{0w7UBND zwYfPpfud+NN0uE7Pj#N`IEiYB?p4hdaN3u8I3SO}$t?WHFN4@itkB=xWG=*2%Ne3x z(Lr6WdH4+pI-^~HLfU3}juIxLKL%{o9Bi$a)_;KL%y2dGJ%Z(;Q~KUkd%!sJf(C!? zHT%N1O~kI3&&dS!s-cs0DCP8)nJg3!#?D5eOh(L0Db3K+=`6*zgVqRZaL`a}-HbiL z!3xKpHk&75yZCQ+@f3@ISMB$ryD~FO>%XH)d6_Nu`{nN~Z~Mdmn%CVH8@2tWdzZ)S z$kk9b;nPUio`nYYLl;e~y{_DfV;O(@a{XR}9Z$sC9|TZp%9SlUOb#-HM*}J)46qBL zOFWts0xS36J9Affxi=IL#}*1YK0mD&2WeEoA&;CS%1@+QoFK*JC2|Ka?*VQaSyq#N<)4X zP2>!TQCsjOT}a+>yaYB?i`(zuJ+Lf|9v*}>8^>fcBI4m!CdOC|oJ8Zyc9_*TJ=px+ z8>_CDE+P>re3}PU+w_5S_uPLTFk&SxbR7CD@jqp7Y>{45*scMY*SHKj6^Uryl+rxT zAEWBMCIp3i`a}8F88i%G-pk8V^4AkeesOkrM58f|;<-8UUm3G(h2+~|NUIrU2b`J| z4CLjAfz>|oTnJ`r8xFedPW|YK$kkNY^~OmYbzxNnNY#b#0Q>sdNHl*s-bQf3^(#fA z>4%7pVk}}ij|1R%=0!Q}@RA74ax(K?AB$SNgxL4>&*WssF#kToVs7a{g;`Wdm+y_! z5*t}{-QRD~)F3E6zaP@R7mPF2}jCcZL_RAgivul42c5yPu*i|9g3I_PJ$NXP$U778RjTHvNcmZOt?|xiXl3g ztee`xorpAo(k^F6peF=i6ET)FG=Y9R@6%2tV4f=!D*4?D8JmA8X!F!)#SsDrUlaDB zDqI86Ly-TunR8WsID>S-@|&8wvc+!m1D>Ds@q7LSWUb$t6%Uj}Kajl@09P&+$-+Y%3NkP_w+}ik6$=6~F_#WH zEf=>vMJ*)>0x~g|@x?9`5IHb73NK7$ZfA68G9WWFG?(BKE*1qdH90gmmodgJD1W+j zSk&LvK1?@ANY~IH-6hi9DM-iA%+TE+4bsxxozmSUUD6>XNO$;--}5|2&-p$7z84n^ z_sV;%z4qRpxn^i6RMeS8&A=u=DKN;HnVp4A0HC0#2DUQ>v9mF&0WDl?jU53TENpCC zC^R(UjzD8)D=DMn$k=F2~i~tNdQob2_UVhuJQY?1`q_{Z@~mm(tzlH=Ybghc2|_t z5Y^C8kz{B6T?2p};0knfvihC&zqrvtm;wG!gSeVIg6;lE0HC*YcD5H_Wp#6NW3g~? za%KTLTCmvL{=-kh(#i?o27h+60YF|IfwsVZ8sh>ogS6Ax68NtGzgGpIU}XvfIRSq= zNrC^Rw1cz~;slX9|C<=35zfD3+WyNN-~A0z2oQI1bo`w|@qbm0|4#G2)WyJ%%Jgi#e1DAH{(HrYK`u_7 zfA!`+%QgjroUELjo&FUO2r##@1^#C5^n1;$K>x@TMU`ZwB-J&T6(EBLVpar0x&vZy zc6a{A`gc502?YTF9~%#Vou3Q91{qRGkeN8x&JM!b3FUYDB&;BPat1qku>QY;Z36EP&1cHUQ880CYFCWc{7+pLzK$ zX8$dQG{VQr9&8UVH@0;G`dFC*AulLiPR6c4fU~0u(8ud<#eX6cc3yy)m8mmi)FImi z<)7%XAagK)|6gJVlmAuy&k)f6vsD=&`_v2!vh@I%0nJfZm4Cp_kddJO|7WNDS1u_R zTU#Y#J0Si4)6@SpG`6#{_4rTI|1x%{cc7Vym9wekzgqh* zxvaA>Wd21#7Pdghviu{`{Jo`YA+rlPf2@8VUjSxyE}s9=LB`V51_X3+0&xAy4G3w( ze=&x1|2JO%tADJTqKe`N#{YM0{;`t;nS#x%Ko$TFZXSTKqoc703L9h=IJmh1UhI&; zHUql;P8jKZpasD)R?%LJInWxB#qje}52!i~Ju1;iB*dLAWUXK@cuV ze-MO=^1l%ugp0}_1mU9k2SK=~{Xq~e>VFV~i^d-W;iCBmL9)O9gCN1I`pW_$v-%4{%CY_nLON^n7lgFS_Adyb z_ZN9~NR59Yb3DJ|A!Z(M*BZ%NDg~r z$c1GKH2>3bF82SCI{wE>{LWwxbhHBhSv@XDf9!2toc@XrkvRO>V!x3K(CME;|B|yq*5t1hV}~rjA3EHS@|}QoR{uGa+`nyr zu77{68#g3~la>3Qx*!alZ2x2A*dh7~(jkM!>+>JJrvG{}`sW^o?A`yxzaL8gK%hI&6lHky?5pTYj3FsHmaWC28JMoPN*GdM^96J){Z|}N~=J@&x z>mr#_udH=77v8OKj=kx%qb;`k5++dx@zPWWwDYg~*Po({dqtPu%o^o|;El;4*!Ohd z>_SHs3O_~C@1_gf;YiZoJFI`xJby&g4R`85C4nOKqb4b{kOt5t=%L`brS&P|C&7kI zVnD|gICZ)NxTUHB916V9Y0#a-KSlACd=M|??2f-x-Yyh!vc`<~{ z`7qtjHPW-nU5=9Lw@8iOq4vouimr9|dPX*RpwBT|T0kvi7$n;Bi8XKB$byrzX&XNe9Tnvw~I$?UVpWf5@e?8m%~$3o>4!vb6PU4guvq(=o8><3(BR$E~%}& z8De%3KZ)iV-cAd1i8*WJQNXnpq7a;jH410q)z}>Ve!MjCK^r4Si)fTAwnE>HVWYTq za(pDZnW3ldh~|}2qrygQq}Fo8(`@|JH`$l$jHg5#Z|;YWPGYe!lYfd(gg^u{Wm>FX zN)|Qo)c6jxyPGA2rj;?a_lH6t8ec@J;|K62=XE*hNqG;`~0tL2m>NJkM z4kXUOL9RTXq1uS6j(S8?@9c3&oCQk!E{%eE2?zV!NZJ|Y1205ucpFnP3Vrz%L>BX2 z*x;Zykb_vZNhtGOMt>ia{tt!xyue82$cYNstEvYFXm|s_M1M*HKvEHf$y=VEvpt~V z)fow?>Ws-k^M}P2dl$5+g$osom2#Rsse5Z9`dmyq=Aw6nh`rNQrXn1&jP_KCiqrSc zXxq0}76oLB1+R`m+U8NsbW41N(o^!0VK|+Bxya&+n@QmGo@64)6g+S;Ycogwx}2JR zx49b!bG*GkoqtpK^A-wTN5S9~qT;mytY#VKVE4uRHFs5lCCf)$;x*etO%uo9_O!Uv z^RQ>8#?BP}HGaGAqSFD;b_-`(0<<5A=65zx7mPtU1$eQ{0*%L?eIC;c#*cT5o@Q%v zl<@+=#Izqz12^L%`AoXG1FHP=9zpb5)5CTd<|ok2ntxnxc2{2cGhs1UqjvOh@#O0y zHKK_f&BihWC`5iGTqgxpZ%VrsZS$EN()a}*ZMR*9(f=@S!WZMp`DT(}qNjYett1MC z_yVLjt$gfH@mcuDW3+w1xC@Xek3rC?{H7N>@U&VrRV3xftx-FjMr!)}E|s(qR_+M) z%IN77?U6STc2^Gpr>41b%NU$iD_$j%Jp#;M#pIs9(aiuax) z>1x$$apV)c3?XM+XIZ*_l39!Y11y2QV$OmDoI8$kMf+|`GWkF?mc7L6g$Bi!buj_5 z3r1#?p#DJ9yU=IO2Dpxr;~*?!s6|&e<+w5EQK!MTO(EZvbLOHVgiTM_#&^kdo5Jcs z-+$5s@zULmEZ*NF$P2}IOA}7@)duf6ze6@O9SZ$q?l?ed!E7Boz?H`9VM5ys=UOtm z5ZukjXyX$w9F}W@4X4;kRx@N^uQsltY&ng>w7m6!bw!`0^W(PC+((`Ot9cc8z;;yv zGt-7dCl^+#yyG5!*8=r>-43VY;q|+|a(|awA^yr+i+Uo8$$&D!Wntp2+J$g5==!f6 z4p~3=_{H?HHcPH?5to6!N|h%EbJRC1&BQvA4?@EQJ5&4m=38Pe(QclZz&%Wk?}Afk z*Ye}4H(n9>A>cacL!{zcGPc$>AXTW3K*gsyr%&0X`MF;!>$;C@wBD4xknBr@kAJ{r zzmiRTQbk&nZre+Um2=EPA@{Z)452IJHnZS^xy~_cs zCyjpU*>?lAl#Bs2kLgc_Bd99B*nh%c-0ZdVc11ov8wh|3PUIAMb(zIhrkl_`4PYD2 zFH1$vym$ookKe~8u8Au!FQ0hOYJ%%6B6Mw}EJjc+jwh>}p~uu38=6%c=uL#D_b4=V z+w*KwtVgXJl(s`EB9XV%gKPToa}8E)L%H<~y*~vqV``Rl<@dz&s|X2e=6~}_IG@jT zIy&Yz-FLRiC9|ENbb1J$@)svyVST#^lBo^3U#fZ4Hq1X0u+KDB%EUJSeHza z{lH@=E_X9{4GnevISA};B52!OUb5VT*G*Amcn=sqjf%yvLvbeR3pzlrYFn zRpdTeZ^aMb%*H2Nw*;bzs;^7!k7x5WrjyRV*rsBV%^ao&X=ua3Uw`s#q=#g%4M`@s zG5FaktJGxc=SR0mQ#QxJAMmG2LK{1cY*8pq_dW`Tlo0ncmg;cse^jdc#H3C-Kp-iw z5kJ)2CeNf?pn=?Y1Vu{=GqU8t*JN=0lLcUMLdxUPlssT$7NlYe!XjIZUuN)Ko&=W)z{KMc zj*{mm8Z?}c&q>8p3TNaQcCin8^QRZ#rH$TQ1gsC3nz>!%>fK-LpUZ|P0#su3RlVv9 z3!k-Ytn*gyK8Nbf_G&y5ZzRsZ2Sm+RD6t(f!z2d9rt)*kb$>+Bg}9cy_fmWDz3nLA zE3>uQ)rwMxd??zutK$zn*zIOT`Z`X@7%jE$$FC&iK zcf-@v$cq2)Hh&VOlDP9kCB_$aj#vl*CB8N(7%@J30Jl`8XdES4r_s| zI)l_86_JFvu^SW&d2q|F1nT#>B8G1!)vXlcQf(ZF7B8-=c<0Ufu;xEq@ zuaVp8Tr(m%Fi06^P%?e)hjIfPRrVvM01BCtcteRe_kZo2-s;~S5ZWKfC+o&Ymw3Yr zS?a&6J!|^nD)`8QK4qIB8(*oB3JC`V5;GA(lZ1$uDJ$Ek)i4s zrFu5PQ2Wy`YhmRk>v9w8)Y9Mun7f>^3GbeflLxQIGaFToh@tplNo##qE_?7~i%i;k zMyd%w8h_Qcq$}&0GszD{bZ{)?K4zXpjT?P-53_$8rSi5SM)Zp8YXwJ#u!*F&spqfl z>2-k&MZ=UI^bZkB<&C?uGIgMdFUL3g{RsPwk#oz;6`REk2%(PO>7simRD-nQk;|-! zz0+4D&n{hR5_&?Q@}B2m4<5$g0f+tq*l#b!M8oO-}@Y3o#g^dy|SQZ<}sd^5d<3WmLt$1f_ z6*s6~ttu6;i^7VJ=TK_A9qb&R-w6jVe1D;*Mrz9ZEHk#LR^jb0GWsGgTJ%ckd=0K` zF7J72?W`cH<7H}-=Vj`PzS%B@(S(f1Td@VYm5(Rb8P!GgtvF|cTA!Y$d!|)J*|U@G zV8dj3x?z@j+kJx1LIbTmce8z91sA{0+V*s(bv)BV4kt8mOl)@aiIvj*y5nV*Wq&Sq z(1n_o)lD5MwQh}*K0akwNS{T*D_0em7Sv)g*pe*gkJZLq=sWm=GB6YbTrW8$<&85< zd5k*L++vOhul#{SPL5rumfbY5lbf@WGNm5`{W8qbKDTKzMMK8!D{yCUw)ug61WspL z1Xo=Vbt#BH)}aO@#o$B~`yqB7c7OY<0jv5@-e>sq)qJ(}FU+v3^lL9oS+TLVD`>Y! z@B2;PlN*&|e!0JUtJ`Taif{790j6m@OPAQp2h1u(NVt;2jP!KM0 zS3-wy6p<`peN4n{h_RXM3$mPHV*2iF4;qM+{DnloJ22=a!%6n^f_|Gb~v_i7^`NJ3axc=iKic?{O%GMh-QqHW-irxtC0L*+{iy|-; z;|e-DFnU+}hZl*u(^RfQrhlRLKi<}tkpJ@VJs6|C4yg!nvV0Tt6r-`zLsf+T!2QuW z!iO*}`b!*i_RuhYsT{!6rZqPIvD~3X3>w;jg!m2htb*iz%W){%9tO;)4TBk#=%`Dl=e)bFFCUO&jC^TM7_PdwklB(d02Np6;i%CEJohz{4Mruc z%}m>LSc~?zW6W03DJ3_a{GTt1@qNh)^Ght+gGz+)vP~>?Re$`hbYfw!3!nTH>)J_E zU<5Tj8a81|ky$CtyJPUP2hv52nqb;ijn|IQC>uxXMeRd(Tnu8-KqW|T@`w&!ka2s7 z!C~sqTZd`9&v4RvP7=Ao;3sH&RzG`eXo*^*>R&7%cu!d{L3(l8ZalwZX%FxifVI^V zamE{;=@dFErGR}o+V#!@vhnu8!z!1v%Ffe+FdBV>3>$;Rn0cbY;y?rd$6bZ#6IYZ z5L5^xFj&n_Y+o@w()+{3=W$37(pT?wzfmWxf45_KcIV^2<^bmzRTj%1&doNn@!9r? z;*5dDCZtJJEw6}y6+U+1&Xtneitm8OGh!oK>Yks`j1zfI`rJ&;!LA{oQ3<^hD<7o= zPNV+y*nfpy=k=W9JvxTgw}2gPu{F`Um0p$CWht{$#YjIN1M&vWHByl&0&ysML*sR? zh3a}EHOuZiT&Wotd#E)?n+2DR$V94R=PiD6XqS<&&IUUgO)yVJ>2kBqVJrDx*VGw! zH%*0=rzZxVdQ@)s^^&@ZA0txN(JM^YaES?%@_(3@+G0gl>1+;VqcN2Ge_E`4%azk* zntxQx?eO)~Y;dcRNQ7d4Ce6idCLxTjqUz`p{>Gg%;fvzBEkfiP8oxbh}`RxAL4mU10IsYgtXMf2?vhiK*QL?n4K{6_{^K5T~fF^`?I(fJXwGC&BxE&k|z2cUdF~6n-^xM8!F`NujR>y<2P2c!;KXo%p7$UO;&6Wq-(l zq1so!s5`r`>i`70`HZFd@QZ|HW24vY6KRYU?OWKH;DENJ*y+g<77ek$J}a({pC*Ty zktiMIA+Waf3BV@KrE7=4AK`gbO(m2$clH_t-zh##y9HdjSLMMZ#KEJD>a*+ICNeG> z3I&;)Y>87Y%8s8ZrF4;MhkfhQ+J6+r85$f&3wc#%g6A#iO~9S9c1^lak*%ImyQ2DG zmmT|uEttM9HQG;0e}kP7ZknJ};@6D@9ySAo{Flr*((mwN9-PAzq~&y_dWM_mShKhy zopF}G$|MaV-p|3|Lr*?5+Mp4JKYYTjil8BhH ztk0TCO--h^7wCyG&feTHm&*4;++1`y1u;@|z%EWIt62LMAoA$omO-;^pQ1T4%^vua z6NO)y&yoCOX-^_{puSEImVXH}CDP-*nR*Lv4(Mg`Vf7Pn)%wXvF3D=TyPMvudDvmu z-+kfLB~2V(P&}#4Xeyp`)Rkot-wuB}*2On4@MZU6D*`-nIYGzn`CSIK!ss0b7sh?~ zr%U%n(xK>hV+4KL(<3gf{cniNIs0Uf_AVYM0%TY+%18y@lQwbXq8700B2_A&GiU0iuGL|v-fWHHu-G;{CtfZg=M|B< z3>q5?LE{<#U|J8FM1O$Rh#TauFKR(To=lEkBocsb4$h#JLGlWwZ~HpAEe3fXN0f6X zkUYC$KZI-#_MqVhlBMwT?9Eho8$2K5zz#M1ONU!=AlGTXZLuFn(bI zj)YncQCDSztr@=wRuZqNw`RXyR5Cv&7a&@s$KAP=Pro{rTZuDK5*r z2IG8sM2l?PpDIM{Uc(R9_;x*rS>>Uo;z`|PAd)3VB-Hc|^s2fh(B>E!LYsow8m z4cipzczSIRN`JykNR<@;r(*9i_^gHK4He5&RCd4Z0t#pct;9EYjp1i2v&nZ`Zk+<)d&lS;?Yg(n`9 zLP%Kh(SJ=%9(}!}rpwf(Om2zXz(XI$Xt8$|U&iTL9r2K%KK?703uYGOs0Jn=VkEW1 z!IcL-qf@J;?j?}6kI`QD`994)IkhlVD+ZbgSy7^BS)G$t0@bdqE$rJpTZ_F(MR?)1 zV6Ers$3$D7qOaT$0-To4)-X;grH}X0ZQfGnkALsO4h71Tu?NMbu3HYp@|JEcwF?gd zn+!i}6ioH?i_I!s??{voPFD26-}qmHdl7jlxSL%zutM|zojMXn2x55e5ovE?*4U-a ztdLYw$MG!5Pu@%ww>2|J%&6LVh$^mWaAGsrWwzb<#l%E zQ8gk4< zHN|znL2r_TKcU@_bJ9zYeyD%RLozFf?SK00ouOidUK93&5Jc@(bJQ(iHOu7gd#sVz z^wk3Gg=(QET$r_h)8M9MsK|snGkSb$re*XtHc%^3)5;jiVg z!TI~5tvUTG>Q1}KAIrF%+f~x`$Ep3&SWFN5oL`}wsv+9_%P!oTR~vBK zt*w$qDO5LiBJFmC@H zC&nWXHVnt7Ry)r)iB<-4&zp0@99ECvpc-LQxj!};IdP}0V60Q#E)Xm;EaL|GOY|EzT;w!>Du2X40N`QyGTL{w<{K;Q@z%gr1#wwdq&A z{vfIj!}&B!2#Utip4RbEm}^k-=`sDA4oBf{@8Z+dE)e;s%xVYU9I%s!^D=n7Y$LF& zq)P@bRT<*oIATSaC>FALBX0_oJ%<~+o!??qPWmGH-822%_E_AVCKrDl>*pEXH{MPX z;vcb}er%ZdVZc0U;(rsoVzz7rt($(9#Zu=J_BH4sC&>BAhz`D)>#&BO4BIxTTQc4g zKU*@M2mT%76uswL_IKpsXFft}>mt2nK^#`nSB*l9ub)*KzFzT1Gc)e*=i?|y@WRoa z62^RN1;g4?MA&|;iXVS(Xz@?zOi>h||U0b}VnGogF3%ZtwFgnYt0&&1`23 zRjr)HmOio;c9RZD?_U@>Y!9gwF~hi4ZvBRtxky0ZPv`w5{nH?8y~+$ygY<5Yk=n8N z$fM3^I!rn5m~(tSEMsqW>U#x+Z>fr{_er0m>knzd$JUT0VrGBBE{vNAE=Jg>N#JH< z+zthQR4tg+F+E=@^5lHgQ4?)%L2HyN#I=ukrR>fl3J{TL`dlyk=uP4n>vIc{$u)p&}j*3A8-g>`9;i3MiHRCj9x-3T@PPA zw$PuoyuNE@vVitlV@HleEpqB?_jt@SvNx^a{;ad*EO@=0Rm6{0%Y5;D`q!c<^%w02 z#u33=HA>32UBLy9rd!_aA={g1joK zM4&Pe0!Ob`!gi6GQa`2vdnB$?TCs&&^C<>EG{Ij76%<@El2=XnL<$gC!U`1hwTKWQfzcw*>Jr7cpaH(?vfll|84HRondPhAn%V^eO1GI7L*`0zwCD z3yy#ENU=LDUnJMB5rCtFL$pVTbp|3;>kX`9WjT*EH4TEQ5Rc+;?@};WIFZ>E~c&zMGoVI@uuO)Os0#jR1bhtreo7TcMA+pW%X7){I ztyId$7U9=o{w+-`hJ{_DA~@3ut|z+B_ld6H;oA? zo=$^qgNY=n(pI(%*nvxW*f^*$Ewfqk4LOl(@#XWHLgBhnkxpRxgi&d#$xM;6g<(Bz zgA+69eEx`JGF8KX!Q;^dEnfpShns(gS%Q5duv+di@@LG4!XLHE+3762gq`TgYzyai znY|+4*T|$6vcLYScT+7l%rx}1)RVZbt{~}tW!s{Fy7Hpaq(FFs9MQglussi+LytNw zLV-GAne1f_X<)r4|LTh%u3MJ~!x3PMCf+od&DisRMM%Wp94ZE1yjE&0i~D~?o0h`t zlgzm2yY(rH)AXN5SX}P@&T%d2VR-W={(Wt_-*^ZQ;oif9RLkdte;PB{&*|K<^RzGI zii@ZYZ9e>d^Yj0DpAOfsM#do*_^Bx~99NPg(D3w>wR|fKO$)DqyVr8C4KSM*wn@eq z=cMU1mA!yfwK&x&ySzM?p3#3I2A1hyRv^wRKhyCbel(Q~T#;4wRFPL#6n(3V#jD0u2q zGUCm4E(t&=6C}b5@qXVY`kh0E%C%VDUaT}kGq~$#!(89JT<+8hx^#bKjJxbW-ziyV zoL5a1A}D-KHtS>{{uOVE8pqPG@!69%&Dw{_{eg9M$q81kGml+OPkF99F4f-0;Ic_X zWVx$Rrsj6($+x_R6Q{^M)=W=YT!TMIi1wMj2OO|i4uKxg1$hMnI0Sq{$XvflThev4 zI~9Lj4>4w?9nP=JB%ps3MINic8tQ;Hb|S6(*i48())<5ov7I%H!+G)Q?hg2x;l}U! z{Mieix>DfX*gBkNe2X|HrIyii}Eyfgt( z{57)vYiF+AxJ#BbxdZFxF&=9^+l?(KX{_g@7)-RcXWe=jY-fM<{TOCQS|S`$NxIvW z=nM2S%T~3{BBN>Ng;FkKi{%&f1rdgFmtGb=1CKtEH5|(1{ARN!Z_g z4F-3S!|oT3o%h6$7QaK~<<+4n&v_s@Zwm^LDf`I5i0!gqgFB2~C1+pO+f^sKAy;$neE7EhlEKzVct|D<9rcC#Y0cQm zOizbUCzy~mRc_%e4r*T)KJI}KL4eoWN+~l_sPYjMt%&HXSN*>5aR#WF*X$F}o+U;t zwgn`3A7y&aw4Op{=~YmOk22ZmRTM68$j_`ts1DsorkE|`eZi;f95v?eK1=;nCrNKy6Dy+RtSy$=0|?zq&Y|4-#}hEK^Oq&Wmn7w zQE%zLBG(8_C1z9#GQ{A~p({OGmG`~wTF)vdQk2j=9ug2~HWqNG^pE>6^KMAb1*=8N zinw&^@}qx)ZR(GZ<}Sx{d1nh5TZ?y3*&*H!dM`_amu%v(FdQOYi&-+#w>vy0#RS)% zkv({E^nSIotXyRAcOK)8W^Or>DuG@@ZM~u|B%mxYRMV!H>E{p=rKEL2P{N?^E+%=M z`t|x1OVnqvVevO*wOI-zRL!)XTWn$!740^&P$qv^q1LwbKlgwgkV6;~$~nJ(vB8d& zp&SY1jeD)Wx4Ne{!*6w(#phMuX_BwnqgWw-S-XOmaiw!RR3nyfv#aKDh@Rll^7{S7of-)PrtBV0q3X^! zT?2o*MM8!=_HA>t<%}atB*j62l9(j9m4x_Z|MpXo=rETex0_#uF*(a>3~TZabH$|T zq-e1>+IyBiE;8<{9qx^RW$zyOn`+-0jzg34?Z}n{^h)Y`qtCXX)GA~Te$ZCvcnfr8 zC4QU!^;WOo2fL@qC$ZbpYHhe*eA-E@5^sNADesyhofQ}!R>rd*Hcr+ZVe?OaaSl;f zEgNQAvty)Eht7y>EMxA|=<+_;A%n7aStlUB4ijv^F4bUU9O-cq;`P*eRV3-CrM;IQrg__|5o zR| z`{BXpjNOuX>38p62P!4RTTrG|uV+N@`Qe_K-NfWBP^#Kkq&R$Ddq>4%zkq)lTj{V8 z=GcVN)iq|Hji^Cxb$Op}ux@7*8?qDfSX{agS133uoOZ%(+`sAHm-m&dXUU6wff5@X zRG`9V0&Z1p7Re}jNN0M<2!Q!Yum0oQi}8U)uPo=S{IY+<9P3wIIT*I=RL3V{C_{g zhVG$nu@2`@s9(z%d?$$m0a|DJ4b)K6Y?myIK$7j$TZpb|)Lw5Cga6$Tqr(I~@nwm~@Q$N%~|$?LQz?y+p$G}={IF+7DDF}#206;5XdSrXCy z#G$0)bMLa`&+)@1k(S6qnFjm)2dCNR?-*7Ql;7Grq1tG9rsYncP7YE!zOB%_nx`_S zcS=TBI{h%Ax!@aHrfVa*DbEg!kLmXJW_o{z8Fj0H>|AXKGu4t; z%)8%g%>!Nol8YVUkt;c9XS@Bh3JQ3&6YeGrF$5PQd=wMIIqyDF>#)KEnJ-DblK42I zXV?Zqki&cjAHx z0oF9`#8^7xTTQ=Hd!iyQKxqx(kWpP*NgeMdKh;Ue_G3l!>_F56V%O+qbRbKY@=4vH zi|7kNh(xeqva&~fPR`SGp^*(oilB_C+g>6eW{%)Ij;yT~#)J{#(e)4=!lZ?-}b*#_6& zpV8$lVW?&VuX%X$Wv)YC42#C$+TuB4wY zOHEU~!~=hi<7kHw6itzkz;IR4Gi)~_UiM=&Z$8neKAS+cA;`tO4L+aCdA&kz#qXv` zbv6LVe!OOT-ORc|cRo8$1uFFj>|P#$>6@>^VnIq|tc?!xL1g-={i?KB6b8Hqy3Ikg z2#C)qx-PiDk=Dq-Qt5X_oOvv|lzZ8M9`Wm=+9Q9k7K1Y)+#-cTZT4#O9R*Pi!S;R! z+AZHj`|4(s8FQ)2vFqLNGSc2xFA+P7m6Js4j^Z0$&N&KDa64u#l})OHu-afQ=qG&o9A6-6tGi3@+-WTnf>t@gU9+5d*npVxs` zn1_G5{Z4>%MP-{qJsutFO*ti2hwFRBwk=%NYG2YO=(`=sedU+k3o0MowPcJo28Fxv zRDz-cOQ&4dwDAjFoiTAsj+5=hG)A_$aiK|@;3o+c`Zj#t<#GpS64zQO3r8YKjM;D1 z$rl(5GiQ-CZt~ot-6}h0Z<2WqvTl={I%w$mf71k=@^K4H@c}*-$M(^ zjZk;Kf;f}QIWiss@}I?{^w`^PqtK0EZNpbpolWPU5tf^Z9|C+*K7NF~t?yGy=rO}S zUs&B_9BTMDju~Z2@#DPj&h!HiP!b26Q8fRdv-=)4?IYGg0}y&W}&icITxT1P3s z5>eFCmj#cmJ$P)0cNQZnXD&F^q&6UW#d-_tS3#I$ckwS4v=N4hdkLSQ!giiZ4}FG< zQ)@bxOS*bsjoXmKm|wdszjfzTM+1N5!%B$~%J={rJVDzzYWHWsGdp*KUu`d)+BtAT zEMJ_z_a#nk6yQhFvgx7(yAn7+@2F_nxjAYVw^Prls30`bR_e>94n#z7ym{^4F}R5W zm2@^H*se_bGko`wWE{zH048_1W@3s?;9D7z!OxmyHMJ|ax8*IBKbPXOf%t!!7+?Lr zNXJrrgH7BS&*g@h62xYrB(w-s7luorUY;kw(BL{~blxl8S`!IDA%~@CK~o$vDk1o( z$O$WApI|VOflEW`OAqU?b#E%8)b9cEeh7_ytrk6lJWyJ7+c~~#xY;M3Shr4zikE%c z$0C(0aj(h$Y|o}WJ~gN=j9hR+>!XEV_*FnTf z=}Wr=S$rjZ)GO1mhFmRUg8t2(1q8-whN*SIW{=Mp4g)OB?Jm!2gi`VKXf!yN2ZcHQ z%g1}!F~F2&?wl4uVt%zg*4ANGWR2;y_b)$ta695;U%y^AC;kFx;O@B_M)-13XAA6X zVf_44w^9n%=i?&gW}kndu1W5cKDPZ;{DkV+b0b-QGrKPQfu5sB*k*CW={S7hZv4166-09j_2aJiOgzjYKTz z|Fouv!-KgtfnYCu7L}Dtm=2V*)ZArXjoWM*2rp+MRb7xzoF8wepS#?L!GSC>F~^(m{fwW zCy8HKGfDEEnb&`l*1Y=#aREzB`N27H_RH&r##_H~D%LYzqqR2$yX>|iPw`4FnpUb4 zB{_}rAHR%L_H%n=YIx1mOyQ>`AM~{ScszK!n5Y{K?6R1^CBouab-{*`+2ZlbgVRN% zP1H%WHI1AG2RPI8UV>% z&inmnQ#;U0z}^NV6h70~jp-%OiyHKo>_4=Kz%<^GgHAEC%^Px&9+>VWDj2f|fce1VmS$@Al_MHgl_OdfvJ zThkkO3qXH)Zq7!+USRZKhTOB`4{OCeU%i@;F2aQJhA2fx!?9qs6%BrtW0mLCMPU|t zzdd4s!9%6n`us19I#t`9gM@1?U9=u^S`2Aw{%_3czH{hK;GfUTDM9NRKE26qt3o06K%Q zvyS>EF5e9Y5^ES^QqNBYDM_|S)V=wJTyv-Osa$n@okGF{9vh^v+WW4dpZ&>}wCy&P zuhoBaqqJ_&N-n_K7dHeqFiYpB(|b-+C`-i8YxkM4FH*-QMmfzQu_2q_g?4^unsOk_ zD*|22mXEJ{ehnTivNQhhJOhD4)CDtyf)4VpRb9JrwHg;0%vNdPf z23oggHd@a=aCMnNO|2O>C=JU?RPY1`BamVWaEp7oSwJeJZISxQWqrM zv?%7-^%dC;eUHP}4hf%I)H=Upxe&*o#jCC`7n)-J~exc5Z(? zQF6dMEtG!Fxt2bG2||aKxn#!GcwgUY5jRr|y!Z3$j7!rO&uPjgRp)n8gc5VqvtGx8N6`3oKd+eRz zfd-k~&NCO`{r#&H#t+aZOnWeulwyC_H%N9Nda`7dofN0~RfTXQ#;(nuP_`>ss)pq_ z0Zo;6)j4L?BJ4l#l2yv4X0O(@xWOUJ7G}>`xm?`Y(w^5A4;AsJJ+e|F?Pm`?pJXxvjc{@9QeK%Wr-0Xa>-t6+@~Bclk3b_|!#?O0ab( zHqC6&T$^2fiAnT_goN8{k!hU1RDTk|Ct(NEKW{{8BJ=!2A3zwv~Ka0&Z3HBkClH4IAm#W{KnrX0%^%+`kG6OfWfS%hG>?qef!5=lqTxI4>=mYlW6 zFANDI3iaESBOWRBG!0hR;QHl`4w-yFvPGY$+V!b3l$@TcReUzC!qk6LS1jq7D%^7B z!&QUXEU-rDzC@(SjFf~cGW2#7N_!W%#>$U`S?s9d)U^zgD6qQTC$4bqZ+nz45Aot2 zr@)r73$Gr=>+kB-M?pM&YEmI-Vjvo^=mxyhp%?l5Y6>NFICCHJkT6*ls(Z@imgn40R_+o`mAqA4#ah zCQw0}E>;E@E(iaLt}Ilc-Nq0B?T74QYSo}%ZpyZ_t?B-&EA(%+pqw-T6bxTD6nNG+ za^7^@nHFiRP^Ev8OA+EXT(w(R`{>_ei|i$>4^&)}<^$q)#)GV5Y$f-sm)u`#3eO95 zs@WZgB*G1~nRK*7hAEi5_+Qu4U07LP`t|J3V^0>vet9mrV%;^$q(8*KdNWRMVLOP< zAb9_M^G2fwe3zufJ5c+)7Z4M}b`G7;X~N`PW{H99z;Ay=kaq9Pl(bK*LDCNe@ArYU znk59yTb!u1{gPm=<+`fzAoiy{tswpv(q=3ozl!nIy75!2~ zlY^9g1EqT8zza3KX7A@u{0U5fE!!0al;!XH!v@WBR2)iPk(+s+pdvpM!{({Q%Zaef z3tBF3F`$3$EM;#)?UM}~@>Ct&YdMion?%@eq=mWUB1;NZkuHR1h;PG7SBsk~4g%Sc zlinqv<9zx4E>KMMws+?#`PQj{BfOk2&2O)w8X=DLk`3B}1h-{8sPCx@jxwak2NT`k z{-oDM<^XC^sp%KL=MGyOv;~j*s+UW`PDHLv56yprSFv6wCO9JU<40i*^eirKm}kwD zCUGliGT)35k$1QrgH3~k10UtyHd`$CMBO!A_@m1j-r`@9Z9EFzM=h*H^ z@G*Nv<~oX~OjOh!1$G+Q?m&_HvoLgq|7CwP?m6U;P$$qUZ$FDnSFi>^l?~*&X^5U> z+(xyuFUNi~U|~jMPAP(!(I{Pmu0>;zKSsX7^*HllK!K(e5~s(z`0AV@_%r=YBQnkR z^yzgwZf~{;Jy}r+tE50vE{DSR!FrSQ5nK0E8wTj>f-gTVoZ)F977~(B_uhy%&MD-< zxaK>{+j@UJQr3)QWzY!O|Mjgz(1u0Af&3(zgsOha z)z+`w!@u^uD@k_}=tKn$UU6rKD+S(HqX4#nQne9*vak~~4jYNC5hRI{h3xbx@ptbQ zX$kA&V6Es*v;)>L!eaDe+FY!^x2kF8N0CB;v@IF4)%jjCZam2?!192O6^MU9*~VeH zx%29UZaGYW(KvcMl#gLv?=0=$KbNr#4K3k#|J5d^@T|)i^gRraw;2x z%)c&?)Smvy+#c9M(4l&ZoSUJ|LM~N(_AdczOKFLO8J2uYlClejlYL6&!^AK9B{ZAr=CTbf=7BY%HPD+YMZ)H)ec zM}dWEy7H;qk`Y+=u-hO_gzA%r%R2%b(C$iP=Q|zF{>%`!du;Q<0pnJX`D`*@*u$SD z$j1@`F0_CV<)3)dcmEFRHkvmi-CeXwmuK}NtLkq>Nq3QIq8PXKpz8oS)A#*$KWLs(7;f{}3 z;&(|F2WV>_@Im{Pbni8sx^eoyz-N*Na(axp_z+Xrv{kuM&;p0bBrUsF-bP4oU?XwHUF;Ku8vYFroC@%zqTZ5hn6T8I~xu6Kq?mnd&zd z4wUN4y5FKQ76pG~^yvROYFCEDfxx?hy0|?=stgS8Tvf|D8?}Gf1!Ic&m_i~fb6LJ9 zJ}sI(BnIU_m}e#G3~zM&8BdmE$yghtC_)eV^&xbPYrSPDQ`4bhx;H|1goLqF+m;h{ zQv;1lK%iY@JVJ-KGSDAWxsLVmf=wfiI`H_Eajptq%UoGp-#)sD*&!`}W zv9z`@c|(YYyjzOJJDw>*Wcwj~%7@>R{$w2o_AjY$z}f%@n=@K%32ZA1_}eHw0tMo- zMoBdVhiKKPy*Xw(NCTaT2|8wngYQlJF^+}}b;=hOogr7_I0}-9L1mc74L-pvr{XAIu3uIb#WQlmP;~LYW4H23jhOhYQLF~ zq`}{yMANhul60SoK&K@QZjWGLOSO|9`%nb3+HN3ttrR94pDjZnU2uDDlT}-oz~=p7gMV6V5) zyd@P-K=d$1^gCxXo4v_%$4q-TJH6e#Gf#oFajoDT1$45`;SM6_3)89KkO{1@KQe#V z^)z{-qR~GT>MPS#UJIUYk9BJ9pdM3P*L9HyM&@?8m-}%+E-O1}D~tjV;{fiBHGq2r z;Vg}80?%?a3GP(?%M*AG`%2{65Av80Ud0+rC}_2>oGI45rSuJr7JJwaD4g61w@e(W z3ng_~(=7so2*$3by<)sr`@xa;jfsDeR#+~QQ*zX-13dBUyW~8 z)1LL`L3hhcIARNv0IgYFf~ht@sh~DaW+~wp2z@H$Hihz{JJegC>2|^Uk2inSEo1Ke z`lg8{I%m!YgTE|9YSZFPQfQxSMnY_f+~zXt#t&1YC`3YaCO%%)xRy;n2>!7A+gDn* zHaB5z|2;|Rz*a8|JUUg=Ct_L7&_`bx)<&4ffbY!l`b*41@aFY}D}@^}y^rfc@9tmx zGfRM5w2O0UUxc27bo!anL!EzVUUCmrUp6j9j~MiQI2NZg1NxY1sWm-^JH%9~ZM7XS zoEXD|)`iM`gEM~h*57IFwC_;oRgw6Nq;v;x%>#ZPrAfg}s}dMSLd>L!hsXjRD$c<$ zr;vBA8-C*SPfJJM^r_vWOe3trgK{`X;25FW~)7;K=o37dhVOF{X;9l25rf zKm~Qza7y`u+uS$))gmUH5X7h(q(!K#@^`ZAkz~z{%yJ*@SVe5H17Q_Kj@v3ri7-P| zcGy!I^w>b+EZrGpgHC?{^4Wu{2|8zIglP8>bNNCJ=_<45RJ%{+C}f`HxMwWQj%GL( zgOrV@bO!HVrSobL#yHf2AEBCCH(9_+h|i7~ywBKbsVfFfpT%`2DsJYOu(>;Jh>`5F zgX0ZlC#%y;g9Wnknub=7EL;=@IE(!4&4ep#zMG&svK4O2?GAr~DNdc0?dm$6kDi%) z@H!U^L%8${Xr@~FyK~Fs`jUkvf8k8tKwD!!8C!@ftuD1gZQjsy8ih<+V|6F`Ao17# z2-!F5D__lM(;z2DUgDvXc{W|4h@3ba|HKpQ)ctryB%O!p^((opq)M2$nCDwLTrsMR zDv3XCp)GO3_ZokC3bte(R=Sc+e@-aoL{SZg{k|o0J_w_uhl~-rn>CTFhNy1iE{eH^ zKYg0hz&R*wJ%Z8p{`+;2q z14+z)tQHH~#40A1WKBfUScd1B#NJRVF@AB6_LFb$o3(5k5Z~?NOOStfoowcIDG@8s zQfS_qPaIpu4hceBfwm>bR{WhuMy&_-lMeg+-7kM=D1hZGdcpzilorL<#G(i;emo#B z2=dXb=0b`1Y{yp|VrVvTAg)Prlf0GfgP);^))4MTyo<#9EINYDK%^^&b~ehH5@5E& zas$XKS<7YvU+$O`}~GpdVwFY$|qc&b$C(2_aKPnx7G}l&37f33+3lE{4BS(YOZhVXice zx{eqr?cnX@FXl>;+n~tLZ?}O_{s-%`Vt86HMbnok+m?oEo=-(tAZN$Z)(J_5pY!&U z3V>Dk7e{(;x&|}yosXhBqFMXnqZxyp6Ec7DJc9U^jY~|?S==1e%5Rl^jee?Cg0CAb z_FAj4dH%fU;u0o1xYBzmCOP9E0?i;LJ~Sta%K#zAVO8O37LOvlMi!wLUs+Avh2Oxg zrlJXwjyyh%cPCF*i%G&R{crZyFL>z5vtL;r(ESoeYF@TLJ2W)qtC{*)i_ya(%Wr>d zP5}_a#yQh4svUeLpfhHD!9u+`iQ9#&DceJ5i)>KZ@L^tCvAu}JEU8EL&8InE5*(WT znB3sHFGK`e5|)g#&ZEVG#_w53D5C^>?%>M6BPH;AUginn9y0pKoj~0OF!=*FH!_A^ zHxM175S)vLR1=&xYR05?g^Cg-ba{VO+n`m+KXMTehwRg7Z!zuBYq2={nhnS&*;jL) z!u02CxaBO>d;=2jdy+nnyq_z;E9o)N+nM`M#;G>xwwqVoOgP$yw36WLO^9O3y__zh3A$59%qJy%F$&tbDI@c}-TbZS;?vkxv^~@}RqX+3%0NsB^&D&(&yK-`#0%?iex}~-4QG9TJ zW4i*ki}DQ|5%&sTH|(z6b{!ob`LXu2q30e^o9JX)YNGu`;+h1D=143!JneT^H_#y$ zn`;l6-nEIx){*BzOYWp%ufKvDxb` zIwtgwpP}tqaUeEZUd1WnRPSgF1XwLLA8PQ5I*Oo>OGP5W6$N9XloZI|^w7AJjOF#0 zUaf;4GQjwLF1D!9L8O1#?2KdjEm6bq?vDHd^?;y|*TUNZQ|S#XgVYX;`PVtgLb;7G{EEPbcF? zZ6_V<0S4giw-*4(WHEqgBx^RwdwI0psg2$zYpFh9bk3C^a{hmG>uqk1WgTvl6)1@_ zB*;{%HdvQXB+bpuD@A^Xio&f2>t5PZCw;^(rlN zQ8ouvi%822cZb{;hrOSf|DkCIG0GC)GiHqp?K)O7RCZM zV0TqXn>JCUQLAw_Nisun^5iBA_Kg&dZSPxX^gV7?{M>)L?ob6FWnl%olvm9%v_wbkV?|ay#yoa zCgS8BPAh+>D~QKQb!hl(`folGg5blKg%?Ri3?OyThQ*hAJZ|-t^sjxY-%cp`;6)x&f8L1x1*kKkoxE|W);T2Au#L+h^xa2wo<9$f6vF2 z6|?d~Ho>48GvZ7CEY!py%3?Ce2oQc8`b+4@^hAH01BvDZ5Uks+PtI(`lqi2Xt(>0Z z-C19ZkR_$BL|A*2vLt4h`Sn@rL~Jd3KxlS(sn#ZHFv7I4BEr91ndX?}R_pt7*XcKZO;en0p& z!1I4zo+B(zQe;Xt$o|p?bj8^?cT|4vzr~p{V^?=~m|=)5j8}%!NdzAT6}K z6d_UB{;HF+Whb1TMO>y>odaBYI)9JBeFjGI2yf)1#)ki4Cy5~q8tjn0ApR&mDO#>A zS=~X@hsi2=gct0?A$ZeOk*q5LuyDuAQ0RY$jmcVhd%_K%rWP2rklQE%^t6)kF*Kkq z6KXnC+~ORhfe_v=!u_nd%Ow#$x$#QNwe>_spHafM)mFF%JU&9IC?xT~>Srh^d!RO( z4ilGnk)$x?L1GqADqmbOqBegQ#vGV^$dM7G>~p@5ks1tWtCzZaeZWDITb0jLav*=% zw=x-h7y5pidy$+m!MJISseN6^Qv$4}An0JfTz2h5kJS5VC%@S{PQl6bp6OL#@wS!w78_OTFs6fO{0Jiq;rs=qm zT-lqHC!ALZ$N`CV6k3EH$w7yZ85n=79M=aVJlx9b?`De-cN~Udnx3ezlTgdbYT21} z9axvU$;O5T?JWXu$PMy>%+*-w^Lq(`oh}tiTLjscImU%Kka2mU%;fUEQ5W0~Adh7{?Op8g2OH_e+YZ z4P~-e<5$O;eIF<=<(OUcsy|4tO!J8hvY9%`=hcd8wBt$f0z@KAwK}x?1MJcCB$g`v ze%*Z>?He%1h#Q%{H2VW~#gcz@MS`~=NcNF?E3bCj`$_PLj#0RI3fd4KxAxsMzL56!6GX=& zK~d72<*lYg6jtzQMu792=WMMu(a4(JpiXkv3O&Cn-(BRWNN4=4c!9cRMjt_Ii=q7a zOO1*N)hB4!&CjgFi=oH~W#xR@_cq&25SMpyMjl(9;v{#YPKf>)2@QsEc!N`xYXLaw zm0GyBB|jc6^0i*I}18&sZf~q{tel!=&YZB|Jj%#|i1qE}!s@EVPk8CZ@FrafUb+97u`AEGJ zSfi+2u88NEo!Y^{sh#t4q>8${f|{I?*fc}jzsW_QAJq&`Iwh8m821u-iCSUahKmjT z*sK(C3I~6O*(`^PbI`gc7&{ zsD*#L1T>cxsNe&yudhz51bF8!3j1?1Wi9R>q<{TNpQs*9dfO<{JPL#w0Mholz`$0} z-leeZm3T_cI`L!_LWvK{{v4r9j!TMKbYbM;Rms;@(16{E>$l%%K?<;{)s+COMH6H3 z<>H;QXD2y*RRNuBDYM?(cdg3;Oa`fiMT>t41a+!l$@=~Sn0 zO#11}4x^E6c9#PFaSDQeUZ=j`{mZq!KB6x7_loMsRZV@Ot4=sz_Kd20WT*V{U~+$n zCfkZGIZ0BwOI6yq;FBY&z!!&eWrZRDYx2W>m)p%X@>&Z%?ra(zB%un9E715j9p+BM zJaVXgVZ-h@3U)I+>)>ePh1_(r-Q-x_EEDMUCu4bAy4q1iD^`q%1nI^o12nQBlcv;; z7(K=y%QT#*{+0BnOVRoV7l$5yC;xxY4K4VXZd)7BJvL&P6Kp=ba@RQrYtE*vj54!q z^@O73%h<|_QE4y598API>!{|w*oDNMq&NY8RBXkxr6dIuIh4sC_p&y>Y->5F-;Qq| zhkoj$8}FE}nr%8sgQ#Q4ukLohb;fh}2vSiJ>bsk#HJ83lE9CK03ETE{p`d^1;AhTw z~gmEdviHWYDe7j(B}r5pqa&+6VmNE%6Ss6((!&MpZ%}hiXP3;PKKJK zxx9c}2uw1N!Xgr)gl}sMq`-f8X0kRjo<@61y3sVSb%1sq7_Zc%Y=s?vMcPD%XvdA4 zYaWunTO;)hYKfs_-#CSFI1>sLq+WgEIix^Pv0zxX&X5E8x>QDxDeNIq$9hT;e))L| z=tenYhNc4YzKz($I6f>`8{D&O`S%i zOERBRG%F7f%rpwt=Tcmgr_<(=JJ7{gM_~xol>*wKmr!MY;sj3e)w_R1G} z_QFPFbpvi;6)$T&h=Z2=1|a_)_+PvjptS zOMC!vKt4{6tnjSU!_kJIrnKL4*}eaaye6s(z{YqxhTyJ#mnVNz!kEqh)}Va6I=JHA zW3MW8I|%zqYb)G>Jdz@0_;bDTRa32_*knC$kuJzh>+5~YFpwJL?}=s(fsrPFC$YTU zSTm42qgfF*H$XKsrVz*rRHrsO?_{IRQi0!Yovz9R+Bo=3sRyb!=>Vw@|FC4}h`01! zoJrD;i0jv+%Ta$)`qwTYjK)tO639k6h+O(3c;y?u9R4%ar-V-B+7z@PE(Csx)w1j= z@)Z7%IN4;;WT*{T^Bx>uFVagF6STB;Q%tFd)F42{3q5$Iy|;*^tO~G`2{J%95xQ1~Xlx-Gt*5JAQz=?mWGUVog>Dw*KrDDu zeBpkyJEAoulFYFPc>aIl&w z zS>Vk!adPJ&!u1{eX}dFe&`hJ3OgYU$07T{A1SWs^LCa`R;N$=ZC}=>#W$NStd-9g+ zSCrvZQ`C5J9|r&vmANxgNd3S_GM))ekL07kG8%xW|27iYih~%8;s{$MVm=VU%~VD} zcj{?-5>!8vw?|bixWB$qYibU>8^Pwpp1n<+1IX;S zU@SA$_7n6Mw{i{?UA>fEQMa}T&@!vA81X=E9L(@qzhbseE@Bh+;PNRrj8e9nJn zvPe1k#>AIxFL#|c#Tf%>?n%}+*O}~7QbM?tE@bo?(tJw26`-Giia4$g@ z=40EoZQHhu8{5u5wr$(ajjbEo$qjF8bF=%tyH#7YRr{?^&(zF^?wRWI^y%kE*phkh z-Avx-c(AM{-%WO>%lVESc2kosMKEJo<~CrGLwzOwIXy`2X7X|A_y`4Vd|tqBO{WJW zw!`s`6A5TKuxe=Wvx{S%p8z-UTnsZq1L!0{Uc5K{>1QcClJRPuu5;=T{KLYo^=Jg{ z-sEhCShZGkgF2}=?)lBg_^mnif?GQk*2(JgXv2)U@uy8({QH7{%C+{4hqwx zFR7{I&|QOj!zQAxnMdu7*cSW`C2-5N-H8XT;K)3_NeFxph(ly0jx29QMo zAA)D8Xgc*+z?%r=gk?52R~UiVQ+c~D=WcHS=1&rsbnIe9-ihkC0xO3uCFWFlujH~D zLT_pzD$B3U6rQkOHvXEDcaWQ#(sGwxi-Ko6^!(a=GtyB?BSQMDK~(e{;@k!!_Xw~P zxBOetO8FLV<0Q@vAAakGiyH(^254>~iuQ?bQ z_Vm7JIz6xnKwN~Sfrr#P9Ovn>W&ZKwoXueuR}yT?BOfVHvBL;ByWO9s_CTj^gVeSb zbv{7%mVT<==Vpb8)Lm6>&-tf9_dL)m`S3~yg5CcbAJ15COVW8t=n zvRd#?W9=wd-lfYA!v-x0K+QeDI4QKi zUr{O;PFxd8LxZyIdjd5nDfDk~_wJQXt0j+RFJXj&{tHNJqUVoDA>z*}CWXQLH zM*t+iLl#CEuub29J3!ph0Z`PH+Iyb7gbPrkd!Q+7Cg4)@o*FANKc=wGaC@1)?E-AH z)w1W7Xw=adwqB6VIvLU9A>?xz~TRQY)LsAAF9@W{_vF|xR*%9Px#V)0oXRX5qe2f}W zM0YQ5IPD5i3qSxh11NN~yl_vbTo`pQLktq@DB4^3&Kjg9`wCF>@Iwm^K*vR@CmEg( ztk)r-qg|JYhgvth#_3iK!jUwnFy4J%qI<+z?Y*S^+oLJW;3Ut{W3<^%T2gcrQpU$K zie|P+XqNgI7^n&wSPHNm5r52Yf478RF>aBf5v#r*gFt?K1B@H&#tnMH){CH6nc(|% z8s6Tgq|)d5N+n@{PJbhQJ}oGJxr&Dswufs&K~)WE|5& zy>djdazdVtBKT9dD3WzSDPzD=A{~d@K7c{CZpKyHxvP~4nsC zt*J>KyR>CQ>QrSVdCdv?2Da@nL3?DGru5qH_bd4fS4cr5NaOaR5)E`qR&FD)Tw@eq zEOrbDG2$xIoW(Q%RnH_DkwLXqgRixH{&LPt(A%@iMU^!?0= ztC;?)e4ievn&y^^9EV$1Ft7O{e<75rx9x0U<3bFU=`i1Gnr*E-8$Z9dVsryLz4oOV zL71b=Lc}VrXj5W2a$stuXbf+0(rV-!l=WHm32-&G7i^|Zahcy2T_l!Ip=-B`^@!b8 zD+JCRF71Wu{0~xL*b3MH*Tr za<8Sv>UAZsWxc4AlAqBonR+X>q;}N*RfOnGVz6mhZ=NO|pc`Z;BasOor}>pwSt@cl z=IP;L&_pH$mAppD1aC_1IB9I-6f1W#4_GjOkd=@&EF=GJStr6X^;YpbG5`oOuu}cS z+n=~7m+tIYqmZN<9l?u{{WU8391Rd|Os75PpQ6s<+lXrl0A-tQY#?<#95mY@Xaw=F zYSe`d(~}qiV(Leb<}{q}3KjDgiC$UUDyLtKlJHE|)ppQiKgIneGIwT4;AR{zB2eUu_c^% z-+6+Bmz^hi9oB_`@1r!ljrcPp51FdZk(!@&A2&$R2g~e z^rrwsDWDaSqw*=jN>7=S0+fnH+LC`ECLf043%e3v(w3o2|1eoRf0y%!$dY7lvQv|_ zc}ID#DvKgG8=e4j4D67|FQerNzC|pj;6I9NQv2$A-W&~2 zR0!UsyWUvd%J9%$ZX~VC!7X4T!8oZxC$fdzF-#s3D#Eg8^8<*D0SZ=4Mac`(^Xl-p zYIR%7$CZg0?{-%ieda~QOuYsvm>Y_vL={{or6&ArHWAjlUK_{ixq~F7%=Wz;PUgrM z(>k9rv0c_E{)D#4c^;9E26=?Varxsyc0Z^P0(&z3fn}3lp$TddpNL<`5?k+uIUgQ( z)w6+9b=1`TT~~H{2e7o6k=&XIq>UZ!=&SM|C0~fxQk?`zjipj4>8|9{I*>x0vzz5d z>yp_}Bf+i4vlosi)d@zLOM$t|1k(GsZ38lo+DGriWtI$BDcc3|dQYVj8r8~Y4`c%ebYB*~h@~F&nM>Cy zx<*mGMqK56?GVKARMzRLV*SHt@j2FH5bNo-1dGm%(`*jFv~Bdu?*53sh}_W`XzKv? zq?-jmwT`7p0V4juG1yAq>n2D3R#5{+)gwOMh=n1&=SZu;HIDI#4y0A#{x1m;ay7u|K+4K zKM@e0k_rcT2w&c0K4KKw+%*Z8BV`#pOZ1+&VN;oC40zcRzFVH=VGbn~KOG}jg1&gA zDB?#XN!885>aECn(hGEDI$;c+ymd%UR>Ygt$haCTRwuoPxzXyy;Q)m}!BC0!R|I{v zcYsRtKntX~%JfhFCHp(0X=EQ%6?F~N#JnONx8vw?VL%O#c$@d}6CDEIhZL@JRWkR3)!N2b~C8~jpGMMp{nux)jJ65)P1z=HB zjboVz(iX1TSo+MLhg54oeva*kBLyd7e;@-KUTh^KDSyPwy+y)gey!e2x6E^9}r)n4&ht1sDSpIF4E_nIU0oA;3jgK zHBsI;0=epO`%!DabI{~DD-}{f8D~a_pjm%b#s&2)+(D>RT|z6>ZyNr5fDG-QdIh8! z75Iw3#JH&TIc$UU3FFTa_)~Yx54k9J=7BM*x8nvp;5QDc8kc3n#(&Tl;ZeJk4$?r} z2#EH%FD`d4a|XU&*r2F19x%Uw2IGYbiVsIxAF;v4k98($39mmUs?M^Ayl}~Q@+ZLS zwgTf^GmK&m3Y&;0&R`S#UHvVvo0rduRK4hRJE_FYH2lmN76!h{DwKwN$?R$U!PPwV zk$~eXw~CVc)Q~aw&5abg6O(ioS=dv&cI&8%x|2cks`8ZJ^ykw%Bz+HXHR(#d!m_O+t0!eK~Q< zNF^h)&ywrb3&8_TaMvhDd4-47U+zFmE`q;z#UuMQ617!-X9t}^UV=5x`<@U38*s>o3uuqKoYvs+~w_+#4H;!{8my?S7?W-@Dbt!f6 z{@*PH!JG{=!qiauDXmTZlWh#E1%P2%q97R=v_VaU&W|VfmI)M*)bt3bkr~cm5CyI| zvs&Ux1zxgQWCsar9PB@=rq>x5JTTHOZk0t>dzhS390oe!(*L;XlZk>~CyTNb9>*bYKeZA=ZsLh8|b*w-K{%1drsY8VK>=-nYT3A78v0 z{Z+Fe+Xm=zU{xH$d4cUn{RN&*ypu^A0K*e2At-f+*`|zZYL#v4t$ zGW+iXMC?4LPA=?O-BtkV$)bCRelqqrv#W0@ERLeh|OLXE*GjwXyfS(YSgbXnEPgQGe6c9(v!}-GF9r7 zmZbKIMJw%((4UjK{p_G+QSgN*(r)hALeTUvOYK9EIVE7mRD^F%d~+inNLpFAUv`;E zgP09h;o2v7?uYw2bHb=zDwgX4wLwLjClSA2_;pn`(8Ge?&RQ~0t5_f;EG28fo^ll8 z+FW;b*}j+(0cb^s20`!Kw8R6DyDSx(8a9BQ!XABC{(@b5HBow9iu1}_$A?- zh+@-6VZXd+>f~I2Rt3{%8dh;v4?O~7BSr#!4w2JnN!p`z@HsAkiOr|KpcN25xed?G zf=+(cTahvJ9=ts##)-A{OBZCke58J}YgsYYNfmQA9^m-;NOsz@J@#ZQ7;7IjP<_|3 z1e*LLGt6h3f*xY%zEodfwpJ`|w_CY2Zbr3okA|^Fgz)gtr>E)r#gKS>hRSXn14YsR z?P80=t-|mZ6k{UA;R&VdB=CpGq-GOa)u)TpE8%zyoPEo*urDM)@RTb)HUSlOaUc|8 z$LaI87GTxsv&GxNUPtBq*QeTezhSo^65Gqd~3d&B&)Q zwu}Tv8@os~SND!DL2WXaml;f4ezh8qXT}}dh#B;aQ^{VWM%b}o)t=Ul7rtjDr1`t5y9e4-j!E*eQ42d)pmHYHY zVtA<-W|TS-%ii$kc^bm>GOw_i%HfT1AW<$W<7c0Q=oG+~^R- zl1OsmpTE~hqOdCOW;ESr{!+yqk88&D7MU#Qyf4@CvT2mozLW5b2Gkar z0EwrP<~1eKUBB+C#Q5+4Ej_S5IVWwlm!HoZM@X3HyNk?rcrIMdatEZ9b4I}3Zcw%g zx1rW>q^9Tx{u6LQ8y*oIV||Go^25^(^Y@GfUNjm3rNNB={Ugd*o$6k%DW1#jiaYi~ z2DO|LLnL{w2#bsE3Xc^JAx1p-1$YE!OyxBFT$--ldTIevk636<+0HsM$Nl?V$_KYP5&>xim24oa>Ea))$AO9>FX-F<@#Ysf8l#x=2_yw(2fg6k%c)aT_ zL)c1BfW@$3AHXhhh>23o^<;iuy%7N0%};EzMh^%jQ=OwL7=v*TP=EU7>P)vLA;fk} z(jtP;BmR@1d5VeXZ2_x1d=mLaP-jx|ax+5EUhe9iN_;kj*fHN20uZ6vbs?sxxK&kb5IREQw zzJT;}Hj?TE$}8|41|d(!YNqdg#b#<;jCq|SIZ;h#Dfbt6h`cJJR9V2T*CIQOrhH!& zO8L5q0?w;uX1Fm5EC3r+|N8H8d$o`s>C&ny{D3Ru%$RpNY)m74LaJOs5{gSyyN^(~ z>ZR(^z`a5ESTwHt29spdLAhmCZ1NysR$!X9hPTsD_+4Iy!CF*0Mq+0cqtU?WSco8X3l6K?~4|}!#X@U z*{J1Y(sz)LLpTv(L$T_gx-gp)7JvazwTD_Zx^qC;KmdCSXp$5yx#L3CH(p1}Q0lK0 z^w^-P-BDvbMHYHvgeLScV}t!m{GLd@SNDYS0TX})%sA4F!|Y2MOC=!}Jl~qFOH`H> zZUQWRHz=Q2UF#eAEOldKkf_;K*w45h(h!N%B+mpy$YDN${cFz1;l)Voz_tF{65K(hd?g*20x^D1l`H}<^&(!T$ zC^c%9qj;@9$2c<4M!#19gz0Fs`ZCTgo7M&PUPUC!HZ%$dZ}&s#+=o&v+>DvYH(?K% ztNhO%XwASQ)anJlvkRP;l}T z>r_L7MPbU_-{`^y#A6v}V-Iwf?9e8$TXE%6JSzONg=|WGbPDuR$KOdte*izZnD<@^ zjR6g3#|;^UKJiORnuSW*iNIG3dcA(22B4>+kN-O=4qXPyz`&j{ro{s?|6)vPy*xDaRcorogkXB+0TGE0RZfL zCRU(6EvE~_!_`+i{OPJh0sjc1!p_&u=nX7x{+$l*&;4t6=?0M^iORN20vzektcQ{7 z{qzSPOx5W=?7XRTj4nUkM?6R6#`jD~# ze~VUIP=Rm2?MY)1lMAEgPBZmT3K;GU8635Jo%W&<@CNEW#Ft*M9nRP&zZaUKFa^>z z^vfLW=_IAk+7oIpCC?yoR_)Pd{=nc|7({*6Vye~FW^Xx2iqH-k2ExA$Fkk_r+$u;L z7c}tGjGj=zD{@j;>wlYFVjrCAJ8Yb*sJNGOgX&oYHz?>>uZY^@KOV>(1sF+oRK~w* zNo7fZna}S`QBe%DFTv!46n*MXl(I|y6n#Pzenqsc+$Ce0>d23}svcTO-ICInm>5Hb z$;Wy`!_2RBqkj)3)(dUyXMDg)dMLFpm4-G*MUNCz+mNJtN-fOx$ckcVCCTL(Ux%;^ zl4@vzD4Pv*>H-AR($KgS0IvobWlHOG($-2`46+Q3MRrQ?tZS7ZRirj4Q1+;?=V9T@ zIHr24QS`&8bmZ3ZmkOGQAoAegq0pQEM6U-jwCL$c4NbTP)=p~2rOtc$y=oA>erp$d zq5ek0zrn4ZK|Woh9dGbq_=-1{HFx1QjT9UDWQ@ZP0O8S-oL?$P0t6n~U3yjY@|q}v z0(TJB4(j>i+kK+N+B%}e`R zxGd-pRp&v+MFYN91DMc0w3dj)HD~4BYDGZa3eHI#)IkEZW=H53qGeJq9`d0MM7S^=|N2vrNRPj$`!}i|cY} z_*Hp!wpJ~zwN-Dm4ki?!sf?cld;E~V(OAPRPc8%>5uFX4RVs0hh==Fr&~X@r?j8hN zA>+}n?HTbmdRDhNAeY-H|JsHux%-?0{Vz%aqmh@ORZpUom>b5cd`KnhOZVhYvXELs zIyrk90GQML{F4f_Vv=2m6uV#zOk`|5T#lxrO7P#?hP%vO>sH(rk`G`k)2#RSouSa3En$}4kt z<_Zt(?J<{=l#*^aTu(A|QKwjj7`Sw#1ee~+0I&{9RIIYhfMXkeQ%(=tqeq!trz^OmD8~zGk?8>8+R-RRmpWqSx4RXW@71cY8<3Utj|r@k76mnxK|hkPuY0qu$)lY~0T z-9irvl@GX>wN61VzyIofUz&cx$^PC{!a8u`9B*+BxYgW zrBtq8wg`kRVX)Db;-p4PpR~wj+Nj?(_6$`x{i7^}e5ryyM!v;f&?muTU5|B40|3B) z>Jw7?1S_vZ+I^uoM{-XI#3!^dXT)f6Q0CGR&(bZg3H6AY{e8EAjFgKK*s9%)pSBAynBw}HQP+{R8Hi8*u$qm_Dizj>>{!Cr;!$F<5PD>8 z%TY24I~JyHH4rm1%lgs?4^H;NhTd$w;ddYbW(Eiqsslj>KW6g#4%eF81gzHtgifvF z#G$YaRwRggMM}fwr&;4cJRA*8N&$M7DgVG+w`nZQEjHvxAPyci`GlRUgxzZ6oE9(1 zZWBwPenPQ7z#5N^WrYE@@HnK+%mR$ft;OM9XoF%sy@!mX96Ebe_I}+h8^2u#wTzVF z|Ocn_KJVKCUPuq^fUVHv}wRXoU;66Px zG8R06YvPfUP=i^OW<#*Ft&TIv(t?Vt9^V{g0)3gGXHBZLrWoKj0=(^138XDmh9lt2 ze3PGRKL#HkGXfAMYHzJ{-dx5yzQeXVKGtm8T3dcLI+%A_b zzWHE8N@ZQqA~j%j8T!ObnF(#`P9x)2AREu^D&y%}61VJppj!B|X|F@-sv;%(Y9HC_ zV1uTn>=#{S2*a>$34jumL{$;dk%s4U;2bo3e#pM*SeENlELj9s>7$5uiu)UXM)r$l zshszE<;7J(G<<3{NN{8n>-J)gz~67_0(UzBoDI%e1OL7>3kxIk%@lYYZLr|L!TT@VTJJMLw2;ABa5XO)u_plJH=dK9eqzmy z$+3pdDn({aiydq3Hp%PGfb5Lmm#6!L1!;bb zm*bdk^-s0AAOE?wAEHlv(FDnQ|3)8EA<;_&7O{hXfLL2^w7~P%j>pw?UV++gUL!`0 zD&s5V*U|uOt+iq6p}25i7P2A~5Uu^^)trFO!fOJSfUsDtZKR-i)?Ld4d7Pq*wt_bcKA7CTd@*N3#6aT zOW#bOUkLbXLv=VnS85aoyLl0FsWfqas2Z^@&3!G?hfVgv?~8*z>hK{^Um-|7Ps zbcX|lyAwBO2jz7oy|P`tbrKod&@Egce`8@S_!#*u*z+t^ZiNMNBcgIApL1bZ_Hz-$ z!JiS{6~KA&Ai3I!v~^Rh6T%ApoGYZ6_^;3rn|_uM%%^DXEj_O-%>X5Y`>i+aU;7~` z(~$sBnlgxJZTp#mNB?W=!$jOk2XX?^O8wy(Yl4@6Hz8-UkD{-5Ub8T#`SuZlu|lgU z!+S4D(wVY`V)80YMg7`%1faTZKYc#t6**1X7C^AX$NU*oaY#nl8+P56fp--(LINq+ zxcWrgTln3O>n!dG3ma30bFKWVv(Sc+q^xhC2*lrNIo4aj{WOA@uC4kitX4^pyLkr z=$A(Zq@aC8-(-=BJMKNd-$xNt?O6AQDRgGghQbtAhv(7Es>_7Gz5MTuL|G=@FMkRcfd7Yvuo8|z~V6c9-s0;z$2^pwOswz zVzC;QIrMQwZsji-3(sK828VGD+?&kW^uy`0rNShpymUVF_~ zxYP9uhUXzSW-Q`13GJQzctR-pLn+$$GAR1>^I(`SnF-lWF;_n*&b#VbK-=S-zDJo^ zR>@nBWg|F*K98l3$pE+yNwaeSob|z6>FQU6T|aN;fJ(Zdk~+ah&9)dO#3^9#UF*26 z`sxU~nJL_{i}pa>s$Ol94s6z;u}Lsnlh3bnr=ILp#u5!t-ded(XSR#_(;ZoL9UbB< zzvv~=#VX!z5O2=}e^omNaFUiBttF!5~b8%KLaq!Li#t0$(Ai7+9+5 zv(%(+{K+$Z&xkR_?K8=gW?%ON?~+arvKS=6|6W*O+N^$|N3wTywB<6(p{f1Sz+QsM zbk2Gnb}7Z-?o(o@G70oLvKnN73zWdf4+tP-5=@M6%BPv!kWiKal-<>Hhsh%DH{#p1 z#*xA9oeIxlGvayl0dA^<-Uhf_Efa+i5RESnWw#ZUx~>N|3jM8fA_8MDk}k(L+Xv#1?QCr+XY(aoQT=APIhaVtH6j zGeBo&PDZX*d-oSEqjtDwBiLI+(J~Cgfq{vj#VT>h?UAKoW2#3L!bVMF{^meN?dL1E~!t3ky^G-3WuGc{)QD zBSQM;2m@5RAsypCn4c^EeS!feqZ#y1$k9G3#AvAv%FM!?42#6g!~w%7 zXJ&8VYDvV(&B2oXmmQc2aH*@Sx+#U>|4;M8NkzBfy1jO3MMR1sGeHg_DN`>B(iF_n zR5ym6(irgRU#{CmjvFkM;k7}tW$l(P;CJ<}g$!&Jdp3k66Nv*=&wPPbPrdAv#?RcY zpZ(X4R6lSUt_~4X1#lr#6@MB{Ds=T{oD$X}@@O$C8&5w{#XJ@~fPhrzR}yH)+*$%J z-7-zG$XtZP*h0hv+3{TC304PbLw-u2gfB&ESuQA#7}Zm9mL+q44L&jipTq)`8y^$y z-$sHA_!}VPqBBaZ$9a0$8_&uDcFPmoJ{AuS?zq2zrsG8-0hFfb!}9*Cl52Gu0g z7)i-RDIv-md|*8pK%!^vc*?OBxP2-@|AA{6T;wot;Vl?wNZPLpE*>ll0{n6-bS57r zN~;zcB+VMZ3ah(RE9Tv5_2o)Qu| zP7E9{r?6xkVjEg%UgBRX!o0&JZc!-|bHKMsq&#G%YcZsG05Z-@mmnU7d%c7zV}ZF< z1U5L{rV{H^FG!YBCT*o42|WxrYWWa4F3~@b?K!#&y(Gyo5XR#`_)tWjPP37ACm8w_ z%tEdWDEkOBXc9V2oM(Z61*{aQ$vG#mXsZUq!vNsKfB}^Y;1VY(7aa(-1ughVi81Ni zM0P2~G_-MY02?-&7YMSbLeY3#k>L|Cx>0G6wSo@fKiR#C4~x@%U#&wsa@BiI$`<)7Pse|?M|UuG z&n|vJsOD0Pr!IPn4Wa)SWaCo~j+zDM^M+ zG%*jM+py&tIqq=xo+y%GDg;II6AC(Y15TsxSRCm^9G~vPe2@Sfh?cXJ4S$MQZd>3RQDxfb9y>ZBP}{a z0oK*dl?9_b5SeLxfvu>sxVTh}Fc1V=-`Ln90w@zL{0l59ci|p<8cL3xBoCgjv>y)6 zj#*)l^jtt0Mn#3_R648wB_hCwDizRC#%5ap1Ll4xthMCva*dT}BQO)bobleZELBNy#))%wk=jW-7)xO1&X0%@zZSHT_9!Z7O8mWdu% z4d9C#Q!g=M>tQ7BbXIft&*BSsjh#Z%(E*KvJE?dIurX zZ45{`X2UIy2R||{gEONK$%-R5gU?-g?H^E?a{vB?e-)%}19u6k1FNuDBeWn@r_dKf zaGR>me}w)~CB6e6SgR|tD1D0n3Mm^Jf7eBNsjs(Icc}lcYTUAzaow_%pQ{#iwg6b$ zyMdhz9$a&={UXP%bm!nTrGJ#4o-dHgJHl>OFw|FNU`0OS$d!{Aw)NTJU&zPPr!q7ef7XN7WIL2HpI#yZ#Y(6K-PfN zF+6~#iI6US%L9)`%V&nY*B-Y?WZB3EqHd6)m6^5w%2)9zyJ*G1~cw z3c@K&D(9d3^1i}lICbt!?gf<_$CDvV#*$IRx(Xo)VvyL_>5fvYUjwth+*+1UxY2nFDc)s1UIPNLL6k;4@juoROEXLK6o4v*sHWA#!T=)XYT6t#h-_JR zV@6ux-OV)15^hlO_+5{jJ9-LZL@G36IC@&wMAT0$TjyZ)r>pm zph+sVLn9ceOW3oqq&}B8)NbV8LiAxm{Uaacfhm>6fs?5)`;IzGm6C}z_ifwpKWba0 zt1efnEp4R}rM(RprvVhV&D5K20+^#6!tm{}nT%xXNfupp*}gu0vQ^nF<88NX;R>@dh?z0=aDq1jt)n{C~jc9YJW%&J9qXb-7n zKV0VtJ3xuNr^>lw)Jkp|fnmi+M06v@Mo5GZ=qNuO?~L};007pQx-3wY$~X3b%uh(j z!byD2Igql}=GxN4dbkt9Cn&o1FGZ(}&A7y&9W&eK;m-(@l1VbI*j{Dl6+e)<+O#1v z#s&gyZ@rUi(~h&%pQeJBx0IgTirP>xC%iS5?wcMp{8oj5PLn~#H93PpmKgacPB<27 zJ^z-aRXN{_0dy|n;M+v7p*-IYjx`7?94KQaxRfLm&p{}pDizEn1_`@f2V{k!Q55m) zhYI=L(JKa2URT!qv5k?=`@sau;ld&hiHeHb>bqW`&&o5J?LswTLBp#glV!d!Nmgx( zEzO&3N)R|LIj?R(iefivw}+H^T}C+AnN&}(Xt+qr0G5X0A1tZx7=pZ&7z|q|Sn{@1 z$NYWx9{mmzuH3p;=_8hNbeD690wKlbLt5Vksnno3R#(o~Px zX&?`0ND5>9!6qCzX_{{A@?(Rm!OYVag1Z3!m-rz~$z@L_WM7XyoLpcdw%-|MJi=FF ztgj4yt&P!V83Nt~oD;V_b33+D7A{~4cz2MRfZ6D~{icpmcq@BNFd}bdM2RCu2WMs) zmfoSbbCVC=9^bLdX6=w)V0h`l)c7m_6UDW~7nMCs_i=;9)eCsuqGbN}VEDBH^K#qq z8Lfetb?~lV>8n5DLrCg{pzZLe?I4Z&NL`ebcJ2>Kx-n~dUDN%Oius`i3_0oii8QMV z0QPuwGu<5AQ(ACn7vxMv^%m6sn~G``C1n*~c;1ds__je^$;Xy(kSE}Go_HX+(`vfBVRrSuf`3YZDBS4-jU5)76VUedUe!h`ZIE z;C4bo(Uf|KM>7-5Fb_4oDaqi|X%cl507@)+Ph6?e=G_Wdl|AYS1GXMY@(&3xL4 zaXL4E@OY=s=V|Y;3&G{eeMf8vhdPQV_FA0$(N2ue8%?QLq1Q5`%=h$CR(vZ`^sz+06{2T z+EI((b#hvY(}gE&uIbm;2;YC*A3<1Z)8(7pmMRJ`zg5Jq`=hnFP0_EAvrqO8E4b=a z9dJBsK~|P61xx{TzSpx}>X;Vg#ZbM!rD^x|&yvDtqMR}_GSLl%#%n%C%}Gi6%pZ?h zsI)#?t$mo^9hhr7kBP2L5eN;w0bgV9&#*2H?YhCpVKoDb*|C=J$KkbsdM45A2{{9H z=^Z|yAL)$w1lRofOdjEwHuVwQZ~oB&kZ3C3OfC#t@f|+kdqW^r0b!0)+t!h3A> zV9}bbi=vQD@B;J1 z$M9`orQ-$(&)y^3m>WM15M}J<#;<>HZ=ihra_RsHNf?cYD}9zi53X&CaC4LRe2Sa3mwd81!7CO%GK8_sa>S7jpmRRZYqrZ;@kta+%%xCO~&AOsBAw;{sQa+OKln? z)+ADl@JqQM3U55AH7R6uZNrLbB_0onCb+M%Mz}ewDOKA5baQNQbF|7CZ{8O@#!N%p zsyCHtq3Q4BEMI8=PT)uQOq6KN^l=hmm7f0UiHm^2NUIi-zG?DyC&Xi}pRCAcf!bVb(XQjP&Cn#)UcV(+Fl~w1XffAu)qMj*UQ1Cww~+?60O* z!P1Al4QKe}^^DKc0-P&?RZz8a?YFVAe1ih8nUP{S_p4V@wUA3-A|+M-yOYP4CNM#Q)#z{Vc{9P!?8}|CQ}yv#`K0N?19&xDqjQ zasQu`pNoq7O;TEH4xU?gU<6l6`Zb?G3fKb{>@pT|HorG}Un5+q*CdSlJkVXF*(}U+us-vsy%UDP5lI0t5cVft z(+C~G0vl*>T(lLqB=C^Tw2$ksVql~dLSWS8&I%V&fkMhbBLQO!U8=GS0TMYs7pluO zJrlx+6NCP>2ark;!cZ5_+$Awgq0r$%DUHOiGOa>!>gBcshfX6q{T)hfrOZSp4Qr{D zM8+0TBSuytU#D=m>7tYy@i+47u!<<7s5OM_hyDQ5*i_(LGd@8+E8};^b&l5x0%6gj zZ6u~qqMf3Zh*i+w+^IK)j56soU=B{(@~?s?&D0}}8nA{>i|=}YAniJKl_2JEfiMy2 zuE(@YxiGzns7!106OUZ1;U~$0b7L0dZ4+z9Sa?_>4}o{p#i4JysNqjG^D5Ng&t7pw zh7Mxd8>=Nb6vQQa%x=Da(K{T_AX4&%qHz_^pyw-8D(^)V;}L-4GQRmjaL+^j9J-sO z6xo{vZ`dqYW6ARZnxL4F1c+pF60^fR- z;rwD~ zqu_%&j}G1tKZ$vB@(AZL9fdQ0TVh3}2Mov2%iA$WYrb!C?#6HKj%`r(v%S4ovHRWu z;_PxZsLx$#vYi$49Bw~6v%~TbY+3c4%$t^CFt={B`P%lrA=RuVT`e-LROGw_C+=)O z^&cUJGjGj#A_cz$o~`fweYONHuCu2#6Nt?9YRZ>~@{dK+0iWG{=bVCwpO5(mLsyl8 zY}x1kzD~Cj6sASLK}*80_W$!uBmQ6FA3J+GUJWBLfSp4hhEdhZ+l+{X{YR$LB+_Lj zVkKh!=~QuWaQ*3IC1N5{fnk)kH+T4n*!~-d5$W=AiwcX1aS3yaii)vvii>ixaf@v5dHsef&Y2Yf2BbV7N-CFC&(9g714~*!f&79K%(|xqSDn8qwAG_5~=@9&vN@;WCp;eYIfw+rL847slX<2RtX57@7Pdac2} z7oa}fYj|X+oCF5X!J?luxgK^1-XYVakj#82ZV7gU~VFvNx%}YJBdw$=Xy4 zyyC1I4QeE3J4U-CG~5sO5x&c>`R_Boba3C~8iu?(ViWX>YBgQB&>|+4FF?u9!J!dH zA>pQI^XwjnL;sJtuqZJJHg?RK!hB00x5LaudJ*r%whZSjinj#vRky9*vy_{;u3Ag5 zerku;lc^Wx?`n9v`u3%#a`rwZN>YNoD_la9gaYO-^}qMiGDvV`hgZm^lGL;M0v#_L zW-isY`&Iu%sY-BfN1RE@(W$BXPdyj8*b&?1ceMSgK|r5|`xNCH8({#mZW^H{_xZb~bnr-cSVbv}^;Mx6A>TIX?h3=5z%`m!@vZU8TdZL+9uFI{C zErOkjLe45}9y}8zlnh;xI*tfH1a2hWyng?6|DSu3{}>AcC%(m~@TSYG*XQLiHn%V` M^a}09}!f(*OVf delta 102265 zcmZUaQ*b3*(5_?Kwl#4kwkLKnv29}qJGSl2#I|jFV%v81`Tle2T>TeM*ShJdwW{l_ zuEvTQ?*(t(?g0-3f&(FdkU%IPG!O;|3xos00}+6TKqMeC5Cw<|L<6D&F@TsrEFd-z z2Z#&A1L6Y-fP_FIATf{xND3rtB6sJ<<7OpeC37^fK@bo?`0>l!!P3o&jE$RvGkuo} zoDR^ewc)VIjpjeARi?B|p`(l8u8_aaSx)>fY>z3SNS4$pa$AEr5Fc zmjL(yyStb3!RyWW4_1G-u_`)&x*z$@G@XFqB-QiAi(&Q&e>SU4Lk^AfwyF#Z3+#_7 z(`n7+^VY~v!PYb@6Z~VUUyJz<3{;{}_kD=ntmCp-)CO{h7$u>tHhZ6JhD-;n&+;pc z7bEsV$lv#{)14L9D{ae!FFC1(;kx$-o~^ms5}AOVTbnC`xdHlMN{n=yN>0aAwK;&s zh_#c=zHUk)Ys`Fe68(1!&VySpwb1)p9=L5W4R0@_Pu%@NzH&_F= z`zX3ntSGV9GoXO(*WVg5-tlSEg9$*XB*~lDb}(rIxN#BWtQ}w2ikZ~PpJ@acU-;8N zSkSF1BNl#FMu__&62g6Hd$Qy1sl#;WCILT`hy(|qc~6Pqa__e(W4k1H;~A1nqSBy@ zT2kkFdKU-$GKQj_y~sY-&q>uDVOgeK<)Mogb!sxqabYeJlVez&;u4 zTku-fm>2wZsSr#|TYjYlj&W|K&Oxw_OY4DE#za*_vT9Yq-9{XX#|~ARd+`xGNcpon%4-m1jU!^ zr|Um{xZ(()jzys~ViQhMHzZ*D#2f?a8~fs-Vy)<<>&PWGlFe=X1=&b;)#OlXe61(W zh;(y)Xb1^%$XiU+3L$&evn65fyTO6igHwA#95g+A)8?cMbU!>t6?5Pf=kz>fkg-c1 zbe!x!j}*S_f#N(?_ye{`5B06G3e%2!zTM3i9`v)A3ehDT+?dR%cM||+=TY8~C|P zy-EEjsBTdPt!ParSoX7|Q&j4H0Qs@Q6gvY30a}Jy_Bem(0+~)0e2$oN0cZ7VUbb zTHKE{K>s1={yduBot^s#&#wroUnCDKkV#-#+Q-wxPXWQ5nHO68 zfBNVLrH%>I_B8MBRc@63RTj>lKLM4Dy$<0|6=m&SPpx$Uj2(n^YpFAr+Ivaj^hky^ zqnE62v@yhoaWjcy?vz9akN=3|Dh=}nKNI$x00icph#!FAv!l_>E>Y-T+ne&WAY}a| zh`|=lDkc7c5f7dU6VQUbx?IgtoRDxw^qr|vZ=*X=AFfP5LJexD`}^~y4GU8sxI!NK z3jKXv_qaPZ1roal7uPDM*2+Ip+6PMDz;)E1h%$#_r~6(Msev$&6GhNgAe-Y_ zHHOOQlYfkP@{UGIJn0ato4=kCr_HRHj%=Gm8uIc7urT!_zkBTnw$@|R#BMuxBdd5) z|H~b(mawzthvWv&CwTN+W>E~XMa#z$pAyU(6Fz|G*EeBt{Qx0Q{ntvdq$wx}KRyra z1{ie85%WF#r!Qk!>h$YJ$KXhVIBwZ81?;g!ZU>tJ16!O(x7 zSA^O~h&A?RA|Ydw#Gc(_Qe6p6I}`8z9vZ_BQ8>-k*m!j#;wD@ z>9Hi^w74{=$6F{B^k)RoiE5Or1TeNFIAuu2~s6KbE*&w@A^HFzF@$?GolSr|lG|~G^P*8ji zN70IrtSUj=Pa`~0G24{%&uzEr`*U(nsVBc*-`8mrHGgD@PEPz=kHuFP);1kKI5YQ;BN5ka|V(oRNYAY=={Te z{1Sy1fa-wTutt1-`>P!U8K7YeK^oe!@w-aOHo3` zdfWYf)YjU&tB&&3A~_17%m^%r3!1$`{#(@@|8U&zJ}YJ}4wjb(Evs1QsO9N(>esPS zTo>r{>DJIj%cQ?Fr1up^IW{JG+%P8;fTY1P$)pmvcB$SXrZrci8hTC%t|gb zU=s@H*v+y1LTbYT*7(`ntakd@9vxMjj;Lrh?dm|Nj9!H_v;L0>YL*AGq5pDmYquAD z>`I0xpJ=3X+ENu2n)<_fJ$9yOZWltUr}=fQG>)j+PC0?Vxx@W*a_%MmV}6xR6{2+B zRn9DDi4E`$LSw|?^yrn*RGM*D!Aq81( zOX}p!YNU1L6~Z+M<%kG5mydm9gAs-y&B+YdZND9g-PZC>cw!cFU5V*hP0@@;bG_~mu9{3Gwz&LHXYv?FT7jsl7OT0fqAl8`7V|ITJQQ5mAYD+fcxml zjgAV>t{Ov-qA-S5byJTEF;6B0O;cvs03^2XHvE)}@u88oFzSla~$ z@>_hpLR4j)%FZpjq9Iwg_M}K}4OKHVI`v{(LErdbGShY-t4)Dqit#Cz7yq674pQ#y zRlus2bJegiSCIpE%{Fm`L$9?ytONFk!uLO_m`OZrGV3q@pXQz>4CjW%6U8~BuZ3y`g`4rT~~*BLH4> zM7Y+mX(^(Z@HL6>=9;x)Fg(%h=F1vbn{N&7<7LVJKzhv5n{veSg9!aNrbGe z!AX%b-;^nd-*NP(gI&P#`?%Z?^PiFRZ`SiJYtTrL;XxuP5H1@03y?WaPWz;3Re`k0 zB%i}4b9fvMtK8)9p<@k)3y}7*wSYNj&L-=f2)!Ylj@mgFI`A@(OR(TR8*MDsVFLD3qo)7`FN_@UM)&UY}s$%C>nc)(FX*S%fU{a z)-3+EDXb=!uk5Jw_)u{cf$T$)p`4}0bLU8|UuR8>qctLvQOs7;cApyUu5R#M@P$27-PiA-<;OEKkf9h1^1nabBzR3 zm<7LILXN#n!CsBCzJakUbvpJFVNN|;%(JIIZ^&pT&v5iv?FZoarB3YY6Zqvt^C(Uw z3uc!HJgLUbMslu6WIgId8W-=zt)5V$u?JaU)ewcAt6Im-0H+@_5CK)TkMBY9Fg=NL zGk+%_ts1`8@U9=(H!WwojQQFOOS$_xfU|Y6yQ;GRaTPH93qx^8#?8Gm_^A=v9LJ+X z1mdrms&UMVC&A;WB;|8LH|rK6F6lgnhM(QbQ)3cTi6o0RBtwWYgni`VrvR^1j=6E)u(3(-co^O!-1SIhfBa z{MpN@Y^`L?5Erd1WrM*FIF1)E#V^qxudm_`BQ}Uy@EFv@1j^vPZ?V}iOXOLczw)_fd^~T?s?{duF(`_h_=%=hAGyHtWs=+mm z>S5j?IQ*e@p5?cbpnv8Wt|8q6+ON~K;^9s(I$~;8y8^@k~>-> zDU1n8FMf`VPT^G~F4cRU5HkeG<>|-f{R;hM%FbX}3mUQEb#ml%ySCn?vTWL&b|c6~ z;3UP>Cj7mfd%+d;hWvZX;o6NQTRHVFMHHow9|#KG0YIh0+t3Fi+>_fj{#~K4>GFED zR_J${-ND;n;qFNM=)1eu{7j4AsoZ&b@z4NVouX29Zz1lE z#b6piDhR)}P)ImSaUfbHoKJsV<9Gu5D}!pQslO3-heya+M7JUTQQ+?gBMIdhI++|7 zWWGBZ%WUNZEtYx7oGE43EbqrA>D<6=KYAs@79j6)54xclZ}Nw`#t)5+rBUzjP*_VFLZMB{B_x1Q0bMJ}x$mMdBp+Aps zJ-gpjJ+n{o4?~R^3|@^4^U-%}uxZ{=xB1B5TsoE7+K6&lc!^S*jLd1@O<%XP@@FoZ z>GZ$IC85|^+1USI24B`$b08eY_CL|WN=QEpYTnU>u*8UmRRm^|K~93xphFc)w)`PU z+>&T4Ke};cSu>U8v`?Xflm(kyo-E%usd2Bg0^%tCq)cL6iD3HWG$3K@jwY&|e@ENp z3h`sy*El314j4-@!dWj%)m5icA~Il|=x$-0$=V;8Q{1O+3V5a8gItHhWfe<^##6&U zA@!J1BoewZK{G%+sK(hgg2HtSB`59&r^F8`4`YW{`o)Ea>P-#~>y6$da5!v5W8{zo zPCuoT0{xvC6E6l_-o-8`M%GLS4Vw$WS6fiVn4sP*FbhLNP^S9>U(Zk?8HfOs%{b41 zrQ;u@FTfc%cB!{890{SAtGZheMEi&e)reLY@H0$V@ zQ@98)_2)1{A}CNF2xQ1nyXG^*FSUnY(PhX~jc)D7$Tr0eX`NJ>K8N-wZZYA0h2;mz z6>+jfXay_WK|ydx2O5Wm@QEoxP1lF5iiVH|BMSjFkYdmT{;_?aF$o)cIvbJ06+c_t zpQkF90>r+WeRtrWLVmiArFn{V+lRT{ZFzbLs@&XYH7g5ir9Vk@pRE@94o?atyn2IL^qmqnaxgAF48QrhuLXqsNv#T4qB0)$ zXw=RDz@N2@p%rnmx`Hp2lC4E46*v~g8X8f+6U_-cV`jMPIn~qngUU=KHVW65(c_*~zCh zfX=<;i86^qmWA0xIKj=&HfbrW*SDh`k4*JzscH56sZwhLwQRMq@J!de;`{9y2nqmj zAyo=9xHs$m<_;O!Sw0ry&aF5hZ0-O#HGbErH#)ShBPsa41=ZSUn`ll}8n?efwh2Q4 zpN5T}vG`!(TQLk+g4T_|p3JpvsTf=V_?m=g&j&Av+EoM#lt4t2FLDFnc=?n+#OG#_ zse=L0c*3H1jd%7)3`K{Ba-m1`!)g4He9abrtTChN@1rp#G9>hRyoeweQU4zPEf z{}#GaeLWGq`&>hH=kuq;Ro*mJ-WYd=x%##ddsHmcS(m~O<1)jcG|Lw=daM~k z?^^tNwF+IqX)tg1AJ^a?r(rwD1rxCN-|d6MgdB?h(e_+S+=*dQ`L%I+&IN=wQ?7&M6dS z=jG;P(4OhSoH})lj{s6fjaksoOO^LB;WiK)N^TjhM!y>%`*UqS|JkBe z_7)_0t4M#HOT4CA>-<6S+2^@z|8{`yPwbx!0C|sIfc$$~w-WpC_$L-g~u}bFJ(y*;Nu=T%}CJIPmO<6c=UgCUGk?i=Oh32sO%>!F?aPw@v}*)JPo4yGG9kLW#aw|3_B^!HG{kv-_13dmzL_2Vl;m543F2--!HA=1Sc{{vZVf+EPZjIOz6fSbI1h^1USFK-{ z#eS8o#_83bn@5Avm*+GgQMJw=zv6L%kgJcMwy@rB9OvSe$!{HX@weL2f$uV9$-M8N zUn{MnMY1fn z{6UYT|8-_iyaxl_Ts@Qb35!YbOI~&5EVZy;Eq=ZxA0twLFE|mu7cR${0i8;`0$PkG z$x)2AwL%~Ede_IiN(V90&i2}11cwB|JK8qdNzF0J&$YC7BpQLKIjS1~mUKkbc>Vvj^3!R4lcPXzu%=f9k-I~1@rF|dlm7?Be^CAh)qha`2hD%b z{s-NE(EkTRD`haF1n579|8cZ#T?k?lL4joB;ZC1_VZlLT{r^*Hb~1KOPF}Wj^mis~ zC@vmOuJk=Fa9Ti9ySe&ry%nw)aky(yvAyeSl>RM!1W(WKg8r?oEv}#c+1y;m->c?T-r^{rh+?8t-`EHkosH?0%%F(;EPO(eW?lV4t3^R6dQ&|o z1na`WBFA7-0^Azp6Y;cNRp=9OP!1Tkv|3qwF zz#%Jiv@$-}H!_BqYvqQJnUh*U~V|4zf zpu;!T&jS|bBC3KC7TN)k7fWO?m_UxX?fy%pbW zG1#e%mD~&MvSIGM^wh2NrwYUmR6Rny!1|s)kaLq$c(2YuPfl~5Fc3snxq${GAoWk4 zU~?l#yi9z)0M55j(V#EJj<)7{NeN-Bd=GV@1GRR$qUP!SlF}C>+VILg83Q3;U z^x$`LteR^U^W;1+XVkmZjqhsw2j67%_ zGKPr>3^goXlN+MrdzlK1PuG@k6csnHI=L9P@81Z?^Q=@u1NbTMJK{U^Czu-rNYpL{ ztiusNNduPqx)XsfLVL$!0zD`D6)R*S2a5d%?#2!hy?g}YYww+xh5>~Bj~)KcU@e2l z;!sOKJhSna?{T0Yq(>w$zFpy6Y3&0*Uwj)4D3EmKFDPh=iF9mw19TeUT9~H)V zcji3H`3{RAod;}v2ZOeFT>w(YhL}O$H=6IgzT<}4^Yhyy5M?$NkItlD;gL-}#od{e zG+(367Ox3_uJ12}k6w&~Pr=CVJ!U$qgLBcrsTqU<<3l5G*1Cm8e^v{V`}ZG)gMiql zUa4{ESr8#8vmA_U4%-T@lhe_34BIMi3Geqla}tjZ0zh^^)K@pIzmf(WvOa}xEiQyQ zr&|H;aPr0WS%Gr8HSv9<{3>_XQol?JBf{aWWgEkg&DoS{wzv3(`6@PqjDj`kk67hp zK27(j-ny)eGgV@0rNNLu?h3);J9dG93)9g2_XAs!^C$5eL!xqJlRFpe!XKlwLip)z zjA3_lct9tlrE>%1(%lz2GlN*Kv{9ss)DIp!QwfNPmJiw_7!B2$BnSVa&q8!{2@M{W zxKxUEuAGKS!)0QhbqcT0SFL014MCsR8F)Ewyw&wgXm z&iW-M>vRPQl{x8al8_vlZPH%m!n}TGzygvdI(FHQiisF`@(mmBFDr|azzQCM z*vnGzR+ILgr(93sEAkue-GQuoa@9BeP%TklTQMonB42xj5tLR3$}cSboRl%Q5QPDz zqW$@AKLuvU1zu&O^UUvga;HqiRpE~8jyHOqoqkO|zx$zce&?iCB(&~mdXXIaVIp!_ ziU5veLL7K`{%ITGd&{I+Q}{BNcF)!FCDvrt}}@V(7D9v zk%z-)xTuU*nwTzfk9mwyHP$20ZBw=nch8TocK?DRzmm<{*w9G7oA_=P$|2&dn&`l7xvatgZ+FO&1|KX3#dIXRw0yv|dU z3(bIdz?Yj~Ag%|fO~3pqHq7Si^uPQw4JU$C0rBE9=_jj6(o6P-&5?IduLzs;fU;1C zvGUI7xRz~jj`OAvyR(6tt4KU<#~rlQ4n`7gJW8n+n}jj$WsP{~cb1OpA?WAELIAMs zcWD=vJvmv^2pp+i)h8Hcr;|cxt%^L*nd4F?xLv2=eYjEbn-^mDoGJN%Q9ad$K+crB z;IQR~EPpNb&wDa7-P~t|rti=VO*g?3hRzEYiB0{h5a=Ww8b#5tk{0KX>wfInUp{{B z!!sc*y+YA8bO!pbCmxuQ_pi*$Gfla&%UeLj_`VQDyZRGBO)^1nvWPIRUO*>$2uh zL86&(F#S0V)J4>2Qe0*%vI;>j%GZVO?)2)#O0lhbgT}0;&<_yC?V;)@4 zPw?AMitQS2VaFewZH5cU{8|ffJrd`hM0+rsUQ0B+EfA{|GFyBIn@Z=)g20Njcc^%g z3X~1H(5}nkGSSVbxfRo$GXNU4c@lx~c;y76U12IoICf6;^+q)Wxqo7Q5ZlZ#2Wr8D zbf5z9gREj(gj7_J3tMEKJQLn!Wj>kTBmEo&PvOqxrb~R!#LzB(Ll!iV17W|YJpHag zC;dM&VgA^C$x(6CdJ1PwMOtPyRD)ePC~dgQdz}UH))N@Atz)FnUcdog0#(+PdXkIV zK%CyV25~3dYh-J$kMAG`az*m8F!2?jS6VGMb@F$2Px=u z6Bb##wTJp8 zsqqHgU0)t{Lk#fFf;6w0Je)eSq5-v3U}p$3awGYEEKIy*J#P2ZS0bmme0;=&^0 zgI`%w9xHSX;lmoj<9*RNr2dzjB4!c3V^qrG;S(d;uU%UA8yb?r8L{f-?4!L`@11y{ zSti$Y`_b}{5|HsH=*N0um-WT$h`_F{!sXv#3r$3n=SMLz=FL)Jd8d^G!Jiz_v<;ly zkaz1`Ie3P0D0%=W`(5WA`R&{fDkntjWDmhLKinZCCB#37sU2Dp`#BTT|4x-jQOiaez2(ykjqE`b%faMtyhq8JbRe zpDiyz`(j052RsSPtU=Muh^;)hJLb#^^k=Sd-_Km-3f5Kp^m#qs-f35RnFT^NyEN_q zoHIQm_n>i-(6=0tB<0#L&Mc7~5LD;lLUR){Go9+T0>)*}y5p*Uy!MpUUba$1iWp%~ z6`}3?CV*`m)md}9ODzFLPu5k}rGvx;<<7#2Wl6i#up>c;u5dBCBQUqVmkmi$Fl|tk zk`Ym;oJ}?*l+MXYoYFN31B%WGkpRSO=S9I>(2bD?LyCWDx0H;%T^3nI*}=EGzHB^g zPlVzi7qinAdtB>uGE<;MG$VFlf<(N`;mKNi0kDl#@eWNU&$By0rBKNoDsS)GH%gUT zy#7Yeb(9&|cET)NoTYW&M-$II&bb!TB87kzKhVoea%}VjZuSKcTC;P990kcIxF592 zd`*`BeS9~Ha=%s#-rXi!7InUKt|(Lf!ZlWd=UIEI0=-HcI93FDf?rD0jR>wq7@NL5 z0#IhiPTOu6>K#HBg3A9o>}rTPruU*BqL_1F{wzjrzKj3E&4%R3*NI`KY^r>zZ<>KU zFf&afm$H!M{3SM)YfH3^F%i!t3@Tn9g{i|p}U56`?RnZQ!W_2>b@qXia<0kDrK zE~E3E8mG%ynMy+DWu)y0m!x+k{U5gUooH|&RQm-xA!hK4b=4Xt!y89;1dR-#`RXyqZU_S2<=O=Wh;#OBM7pI)WK``5yH6o< zy?QkqnY8x*TZ)Kmlq9zM;@e8O0>ISvs*Y?3HLhj0aT!wW{?)2hmoIZt)vh-nPUNCG zZ`Oq(=@!U_eZNhorgawEB09mE{A=-3V<}cW@GRl9^eyQXQ|(YeXv##($mzr>a&jhUy(lGAc#+%4tBLDp9~(z|LL6n^>g_#{l2P8 zziB1+4~*%*tzRRn6P+%sJN6@a-`g!%x7!dEJzS-_w?#J2!O<{j;A*Nj^f&Yd;iKm{ zwrwtpCDNCL(bi zh+o2yyb1%Nnh^1%ZiJVSc`2C(vlHSjp(9-YqaM!H{Bjng-M%%NdFY+?(py?eQ(EK6 z(&8EGqD_A-Pqb5cryECmNhy?>o;Y@jCDan>01E{D=#$byMqUzKBLF>b9x+_*xt5J5 z8C)0&Eg2KyBYfmIK~E4#!{Jm-x7g=Wf2p*ah2J>`B5H>5Xg#Zk{eN!&%;2-M}0(`C?~FWoWak6^!eu?p8_V zxj23k%W?QA>~-(MZSd^SAEBhZlO^1;T|$fXWZ#z5o707XGb3+|TgSh+^0*P3#@4R- z+GA^DY)2Y+sTliYE58E&80v?E zvSE9hS%lExEy^beI@L7pIljNEr&8I$zt=fE>~Lp@M(@Z<0Q~tq>$A9koJocA42Qd9 zhd~GOgZli<08g+PE!!fvIU%T;bLx&{B1meGj(fZw6k2GP9{aN<~NQrz~ ziOrhl?>b_+LVz20uuwidA84pO+n%o|#o}y}w*DS^NIpIY19yP@EtZ8(U;OUcB<|Ipksi3B$gt@Upxm2@`qZpnaAZX8ul zLrMA5+uwL}PZA_Atb%KNl7&4wheKcpWznd6RRoZmuD5q$_K-DFs2yj2)e0N#a#Y`# zuH$+>>!Gcf(6!wzqxoC(68}Z54Q(t>>qBfH$^}1#-d}eXb2m-^!fS#Mk$acyX z8Uxs$>(t=2J+g$jJFA)&RV=-8Yu>zyDjMA?<`0$3^dHKpi(2l5*%pc3;5Az6#}^WJ zKSPbYP;pRF@ZO;qTtn6=1a*-<95NJXEUSen;O=?Liw_(lgykscX#4&#B8woJXcIVa z+hC_%~+>|apgz?9yE4ICGyG&>3qS? zJgNVJ#(4r1>G=~WgNSJRi$id{Xo-mgzKPwqvyn1_@I}jwWw;5JBK)Qj5@~-2S9CkF zzg8Zmp}eovOP76$#$_y)o7+#z$pN)3!(~wbs1LrJSgVJxb+`&sH>7Kv1uW-al?1%5 z!I25!%>G#2h6syj6Q5Qm%Bs{OC-CZ+f>_S!R)1usZBTdxOG*vYHDD%baui<$&umz0 z5-t5OwrBnF-+}h&5o$@WDS(A})8AVbX1bFRaJ&$nVlW%hN0x5cVW2-2kmWZu;aV(1 zZc0WF#s5V*Dd&cF%G#6WV6FY<;2huu-V+hhJ0Z|7Zzr(yg z)Q(2LRqu{~HlU1Qr6pg!Gy(A}!Z}C&8iE5kMYKj7o2Z~pg@Gv%w|dXsmI3Y}f$6S2 z>78geBLEU|BmQMr|NQTX-X{PRC7;eW)F+DA!e1NjzcposDLW}hOkWS-g|jAD>Y7Pb`|Y$jG0gDw_je$N4bkv)!e&f(ORMRNjB2O`!m7_{3?x_Xg#| zp7~;%ZH5MNX`ZcuYSZQ}-8yjZNtP{Gp(}E10Y~7ydvAt{`_h%vLT!$o$sY>Ibj6NV zb{eR27k;N{;P^_a z0!GV+EN8+`d{;75LpT6v2mZLFk%BQCRG9@5MGfyP0P_> zf*`r{usByarNlH0M2DZe-{RCwuW6aW$FSf*+&{-Hy)4mxg_s>p54+e=d9mHX>FPYTJ#!wfnfh)1jeiLGz$7&LCy=!E=Y%I~GY`Z@E|Vp6EP+_IJ}? zqAbigh5ZZit6`^z&A)$jAEM%Z8w0)zXjH>_d{|uc&L5c^Rl-cnar{a0pz1|uJb^G= zO*fhF8Q{~+>90y3wTB^7d9r+CXSJ53c!7o$kCTj(tqp z2A*i{rdCf&Bg{IflN9(!b%Fk>5JD_pu!7LSra@d z5-cjBw>Y8}P%cX0?2C~xg=dfuglE)8hO{$QEwY}|*774b#}MC3-QMI-VGs43LsY)p zAx+@CHYEU`)brPLN0OT(%jvAWE0JBr&a{D}2T>(JCG@y#d|ZEKMn>1v0%2a;_B;r>TTCrNJr~~FH9$BnQft)j&jiCU_!SCIyA@*b$nx&H7XiX&+t7>W3qoNB2tX@aX#ClkL9ey(Xfx^%ivcZZo#>9Die zN}{i^4;zL5!P|@=_|D(FC7%8gc+{*TZZ$F4_qpwfh3q8r5RULCO3+a1<8xdQQaA^o z*gjT+(Tuc#MpZn&lP@WXAk6=!Wc!v*9C%6Z=UgthTC;B{*0YO87Li2r0+2&{hNl6k zN?=@KlXDYe;wre1;7n{8NDN45y5Vy!{5)9Kz`pE!9lXmHLQL!;{kC;B`viAb=&V`( z83nxUMRlc7Yo_X-d-8*0c61B7`TO6e!*%POmbtp>v_~aE!=$hH3BYC?Jg8;1z8iY}>n{cGi*lS&n+!~Q8Se>r^5I_#508uM}EFGK}G z+=hnN&YDFO`kK!3?Xfu@cAiB{>F<(&$S1~{!-D1jCdX#CX(^$zATiGcSE$wQqURC_ zeJo&y6;s|!%&mWPa<++D9|am9Bas1PNTI81W0b|-;Y|l7yEK@sO)5}>HPL%@4S2gk zAeL0U#n9%;?h;}Qp_~AF?xU$k+0NG-s(#tl8;vN7PHSG3xZ!Gz@ zj=hKB?Yn9`BC<6{J~YNzdy~@Us4nCxelpjmsjb&VD5G!mP$ows3i(CCUlpPPD}Z>K zh`|zbx^{V?Ny7r2OhyvpjKsLUT5+721%Wo)D3(yKVz@r5t38A`=OeWAj2M1Bi#uBH zwpFA~qJQsgB$bBV{LK4b{Nt7xCUNrMbL4I5DaoO2uh}`sXLu?x5bgUbxC?sFvOaFF zBtmvi{c9*~DPo@(6k4LPO|7nwz)5B}@yRQA+`Ls#^#ZnQc_$2hK+j}~_hcrEG7}L% z`_*HZFq+DcC#={~f-`n=M5f9Hd?gK!L&hom;|3d6K&)&}JBI_Hr*pkWP)olOCp*&I zaT9P{eAEfNKy14r!ZI_>sD0f2V})x?Zxi&){f#5(#vqf$HlAT(;b!ASkOKDlc0a}d zg1^$AjnBBuISm#gDoE35K zR1;N*LQqbSHnRoFBQ7PQ9E)MONG_3<)mpHEn=QQO&t=6d;woM|p zJx*Wn3>N{*TDUS5TpeUj7RTDz!=51ni~g1r^tdePi$qySJl|Lw@E>5$tcgwS8b;P^ zPwghfb=6Z^A8PEsEPtYyeJm70Qq@2XL~v2i#YkPp#fy_M(?_Y?!&o_g^~0RrI*cSb z#a}C4#w+(2#&wEMHKwo?qM(Q!WyiQGE~^12F=GJ7i>us=v@n>e>Mah(61j#z{bxDu z!T=a0u4-a#rAz5R7iRY!%BiK18@06{H%iwa+UZ36yg&_WDtR`egu1&e*90a4IxFl; zS_HFS40+8@RK(f~wk4UZc2BQqxy|55DgfCEs z_ynLw=an?XrM?78T?mf?uW{qm?l0rFN5{b81XwMhI3>RM)mpVVUsX~;AP!t%L!KVR z8Xa*F#Tw?qLGePuoA+QsoCJq$m7u&{ejOjtF*D*NY}RCP}!{ zCiVr@pVwQ=`Ms5?7i;8L#G1>;*n#Ay6JtPvpQBRU8p$txGFodgclKgLuJ=e#A^d|c zpLNy|aX{OH!sNA!%Gm_+)@)b|t+QUAyN$q4x8GmZtaJ`sN$#j@-jV@6%K|K=nZYg! z$1+xvE;gG~Z;$cmG&l*dLd#-V*hY8JwQHA4`rkZ3ICCs6XeOS1W|3FuEbQ6+b}RsQ z2Z*A8C7=HTGeFG003|&}@diMFAHymzx9>-*K_NG@mpo0y*Ho|29&%S)cc+}YjJ=9@ z33a%c)E;IN#*^=z#!qs1+9$Es*AyUC4BVWovSggSjU-6Gy z+Xv5S$4RjTKZl(?#aMa;=l3Wb;USwh41!z>e0wdH%wV#`FQ51#{`}qc>!G8c4S+9p2tBdK z!n_G=`;pIWC52$}?0t+lcsb}AH|c+mR`ULMu$UWJf-Thx;{$d*qfAT9Awfp$+~0ci z2Qqfkv~q)Vgb&7dp~&&v3Ndxw5M)=1@OGu)NX zJS>Tpra|ZzU&R8?gX8!Qw`I^t_lbo-B%)APn0R9Zs78*1z8CT}d8Q3VC}(_e|cbu|0TdXh$rBX!PV|GtM4&4yPIfOFzhhQ>`m4# zKDaU+xx#MJdp-C|mWRWeq=;hx7e*LZu)iICwyLb4TXBxN8pd)!9S>buhbZbQTgn9x zR!8f=6tNCFUzm_$crD_+TV-ReuMwD3SA!AfsF@)^|I}JV<#xbUC0ReXe}bcOdkay< zMAz+IHor?U{6$FwO{*?eK7?Q6X!A&bgv2cX-Q;zyuGNe|?C1H&(t;LA&H8u52C;e7 z7G#d`loVU@4!wax$*kP)_9m2W4b4*m-TCB3K%p@U2l6X8)xn;T7+(;+n+U{DPkTX4 zUsV;4+rXsLqr4bV*K@SZz6Cn( zg|2*&?XWZ{^6o6{B3RZ7m*8eRtKe(6Ay;L-3it5LLg4NKKVIHGBR~wj$(;MF1y80} zXd~mPuHTLiBk+v2t8zWFrH=8H$Rt}?Wd&;l=VOlSZ{e^<1ifafL2*EOi7 zH-t|C=T8JtD!$!=@Tp%0ycI20#O`FUWNx(%Ul#+CP3ZZgJ%y~I>c>VD!*`5~fzuV8 zW|j>Pkk75v&C&z@vPCB3pM!Nss-M^kS*-g>>HKpbodnNGuu}jyDcf1L>8TUsFiJ5p zcto__+Z->g?M1Uge@lo0!dQc7^VQjJNA4NNWRF$h>)4WZm2CR+9#A3yFX)TWyc2pd zi6o@d1*$RFZ0K8=_6}^X7?>!9ev-(P&a(AQW|DiYF!RF*p>Q{3T^oYXEwPrSH>>Da zY4lvTjSo3U*_H;MSpuLRK?G*xn}y>&`5P7oHJ16X)rS+Oe@gZeGn_kiUy}43ujYyD z@z{56ut(6*wbcKlyPJ)$O<)q$`4V^ClnVOISQkRle|h|F#8B0xs#H zzBlL%v{;U6Oei;D%4611Gnxq)8{pFr1iIT2ht@{ik(mOK6Yo{~aJCSInCpKOxF*x+ zo+$Er^u)vlf4if&kN!lT4l%M{c!K{TqOQ(Tcj8XA_I&f6YEdhKc3k3Jcm%ndTGAKW*3D z(mJg0MS8fSo?KQ2F$0z_@37T4kIt22SK#wEWk>ZB7}Is8F4U0=>kY4KnW`Nt28iW# zZTgF9H?BgdCD`Kx-bkpJStb`>5yU7ER(iBDI(x)^G|JoY(n$|6cf5|XZco)BG*UDV5xVR@~$xlvOTu0{04Zl1s6ay)Db&XS7#R0U9W9|}im1pF{ z$?#l|RE1~GBEzsP0a^n|?Z;oacrFdcLDp)K5^^JV)Y)#aNkv!U@}%_Cc=AS9-enoI z@qZSyEMK1*KYh0M8K^{+0`*HfL}6zD@j!n#e-$jyAh=S{h*-;Trd>i`p+@}smXD=ChSwa|LVeszwBo#p;pk*a& zyA}#c8Uv`luv6pP&&1!4sDcYzcTZA#NLx0*ym=5<`Q>y=*xulb*T3;WL@X? ze#Z|gw?VU?PnbGBg)O6Zz)2L!DV_04Mm!dtCdm;{zx2~Y?`e^g~j z@(~R`jbQmsx>Hqz38cQ?Z`cA;s9>a3oJG#1>3WjUdXK~0Fte}K`?~ED(DV-j9lflz z$&D2iH~9ro2TCN|`>rr|$|PGuY-{h`BE<;D6J0HP$k!~E(_}+d38fjc#5U^CYFmXI z1C)goH?DA$spGnT2<#K~Y&)21e{mFU)8WWt_NANmH=q3Ku|t|<+8%r{AAKpr+^g7` zY$6zxH_He(s&`bj>K}I_M5!X}@hdo$pwjQpqAr;R0mT{wG%}_ca!99dAf(H7n82K{)AQdz1vwFo<&&h?!SAP1&Be$vabVigECp&0B1KKZ%O{z`L9CO|WW$`R1_vA)6aZ4O& z28Co-D23wjoSdIe6No=;e~oHiXXKVItX%y=`8F}u#dqZnL2Zz!xdyIBvv72Jd6JXt zcCRL*%s@?SLKRs&|Z9owm|c@EN=>2vQ=e=B6qkrIp3M7m5* z9W*DRC&X?owGo{wY|5Wyb^xw{#cEJA($X5%#;scxMb6TZLH--lAnY=ILc6_#Mr+$x zEyO$F8~SybuNVEG{fU6MpkCf1hl7#R^4~ggQ;ene2fdUD_R@PK%n8U1qnlrdMI3pd zBoqsRe!X^n-4r&_f0U3yiM5a55N&E#|D0}mh1gg9b!sk}C+J%1r!zl6M4P>gqjnJw zK$e7zR13muBt+W+Vj~p)7xdcTMd9q%T3l{)1O0R4YR^W$@EY3++>I1>+y%FEoHFZR z(u1SxyPooRNe??N!&MfsQI#y$4u=`ve)~NbcV!h0?{C@|fAxKsdIz-~*bVVJ@NQ0i zx7ml-)%T?sQNFnQ41;3d;3AwL=7QQKz^xocr;AcHN81y>D?vnno_cF3q&}ENv?+1M zI7JpeS z$~IuflEpaBe;O7!g~y3&Vvqvm1&tAvZOo#ri7}d#;T15dGcJstcxHzzvjx)J^kWo#Ca!E`yfaggt{6cA z+6((ROk6Kx^g+1X8=NdkCEJskXuu>^qKC=jM7g2{f5{s0CYW33zJ63X+7T{suw~fD zTloGLObz(FDhqFe>H1os+td$Ce2BX(qn_a`m3D0zUPB(s==MnM{Qx}qdfQ-Au7Hr3 zokoQIFKE-E%}q00$^lu9O2pGtd9pY;_=AG^fm=LoecyG66MUQO#wu!Q+PdxJFpA}Q zPviAYe}64u;qFDFPcXR?ghjPisd%$vrjq`=YAK1wW-phRS9T7ozb(bU)f#1W&m1_{ ztt;Wo#(OZ)d(r~_)P-4BB{cgM%%cQm2MadpQWA*ei-p@;DI31!OQVnCOH{;mDW>;A zRDPY@3N5ML8Va0X=aZG#UrwIV$AjF2gO~Qu|ysH?IceZN!E0I=G%{! zvXyabhf@vMSsE7>ux$yZ2V+JMshNkcD1eFCWCyz;od$^7!>CZ`wLwBvsbbY74UEl8z z>*#fShMl`Avx-oK<;KT6H|?)MBh3T ziv&4;ny%Q(csj4c>XzScUGTI|e+wRNx~1vKo}4(6Jqt05^1sec1Y}9O#76~lpjV<4 zR$12Tq1g40ImfA`(?;pZo^>~(k)z|&B+VIT#Ov5%-#cyc_MJzI8d!$1x^#_q^Edz6cg@>i(~2yGIx^ zhTA|!npbj@1}v5~6w%Qu_UrhE4tuh+-s6D?g~a&&q3>Km@jbY*4Ck7%>N$Cl6~~SVble0wy)zKD zNj;MuoMg{JuKG%IUVS*z0L*#m-mtKw{--?U^-LTs`aFf)GfvB zp08yBp5lNgNF#}se}roq&0!0(WWsjy<0K_(WVM>boOXNs6iRo$hH4ah4MVF;KM-&A34 z5D02h{GG?Pwd@!&g}aUZy(83Z*?8%T18jUw482T`ov&@Jf2Fj16pC3*4_m;*_gQot zCThEIOnMMWKxdk7is{bNo6#LBSpJYr;Ph0LTh^Y4nwm&;MvxXK=sjxpPnfT@-gMB1 z%LuiTMmn)2wMdp>QRpar(CZe^p|I04%&hgu&he_c(}C|lU>~x}dJd*9eYcnVE@%!> z*!+6GDtoCEf5^@#nO%PF#vc}4W&h$KCDZHlRB(sZb^C6)*83oJFc3(`HZj@PPkM1; zM)6Z6*+L<5<@HHpZB6EK?)tU!rb|yi6p{`>=V%e{I%#x^f=j@MSN_MLFx3QJ({^ zHMv1BU{WomO^!k23YX<&X-^{PX8EXM&D11qMhKZj^O)^m=eU*9aKO!Uvw161=3rNh&P9yG3UWvdfG2 zxj!;$pK^_i%$PVv=zZbLm;9evri z&k#~o^s@N8B3RL)Twi-Of2FT;2LTrxiZoV|38Mzgg+b(6C#LqD`b;flgg6 z-1!QlRWMOZA?;`0Xz6S+pxxgd9|)p>e_T9ko+K=}GN%P+8)o>co`0H~UIG1~YP2IM zD;2*;_n|q|0Mysy_G^!Z{@*x%<^Kl}lBCk$Sqt z#p~*Y00lNr>|o}~x1&5?2TBuSJv6Dj`n}3j`olOb=dPXCTp@fiUv1JRX&U|we@5qZ z=J^`u?(%Qig9xMRl7n6aItjs05?26c|4LOuqRE;!V8#;YQjb%l5^UW!zqgUK4OricG6p@@KO5zIO#VPiBWKY3Rf{JPh=) zEMD(tsO%q9<2@~j<8sgs%(rhtfBoIkzC|H~S&y-0huY$-pHse?_|SCH$Rm#dBu+n~c;rQv$gbebBt9ZLE{o%$o@@PM{#V z91Wujwcof|+0kZCL}TLz7HTeoQdbwen!m&vJf=BeY+PWR;3{)Lms)?G4wGeA*fJH! zrY0O!Hg^$ot*<4fRaF}+OKo^yCYY;F1(oa!U-~k}*sN`{S$W4lu7LGZlvO(YN0>JRWuD!p zey{0{>Yk_Zp;TPOe+G)Z@aayTGmTHJyoz#JP`V5)`_qW)9ErM{fT$2mF*^0^dF2~= zJC=+0ql#w6JCcg9?52LPLZ5g@E5z2~x#{)rfqnIku>P*8BRHEgfN8NxmBPf5`wGFN zYmUs&43FT1!zHvr@G7;Mq6!YQ8sqWFNnB{199@$0zUGo*f3Mtk{rq~#HvWv7;8YHm z4^k0;Vl4l#F~TLm_r0sLyy8Y2`$1VhgV7W%ha+-`_dXH>38o2>V7K@B_w0l80DZt{ zX5(>?@|hE*hhun!lzTtzC^qavyMxeZrEWJ{8%@DD28Pp@R_G&B2SU;=iq~&uGnH3^ zO@W@EAa6F_f0<{~@$ZtBbZheCE8mEEWDa6;S4rdQ-xgh*u3OXaq0GJte5K#WhjCE) zq`Qz>{0L0kpZUeut=(LcbbvF2sW(g@*xAL|^r9qX2s^m((pcXA`T z{s^mvM7&55ejki?5^t}$o+UyL4Q8}N6&gP%V|tyxS^I^tlB7q#m#d88p~R~Mit=57 zD>FEJf9_h19{3xBDokh;*koOqu^OWUywd06AsHo?H6)32fV+xpj{%yR`F1lQB|-ck zoR`W`rsZ;e5B94ll;jQM4VUHjPm^eKVZy`)dt!#!>l-X8z+?rs!9tR|lmR zW<<>0%I5Z=zT_QfKn!*TikLWP+nyT#SNf)dPbBiVz3LW;TQNV6k+Gu;p3oLzcBCoA zj#RmuRHi=PK(f%zU$7{eY8;w{t)L+Zf43Jy_MQ&XR3P??$%RbZ?WE)vu+)z&LC!kn zvX_NZ+(Ykw8MyNsg!)w_#*w@IJButEH1!%*)=A}e{q|eh62QB4LD(7Uvw~vZ%O|NM zj|tK74P(i|6;vlh0wQMdULTqgZRRvfDsQ8hCuN)m{lkk6B0f6NAQ z!^3$PV%I6QWUm2uwL+&%)e|p;itw^1dW`JazvMwc?x4tq-|3*0?))Y2Js^Y=Q z(z8Jm=HlxDV*2R^I+H-V9+1)KASHlOG?F8|$~DSZ&$mZVvY@n+S#=V7J(I-iWbMyH!bJ%qi|ifK+wN`=_uUe>OXnU7k0G zxgZvv*A6GBjo~pZknWxxI)c)_&r&A(c4g+yAdAWSB}Vp-8k-==unTf3&GZk}t(EN0 z3Zc#yLL@;7js=dsP_03!>cS;E5IB@{BR@(crY;v|3yyw?yvexiCU12+wIKWsD}E$d zsA2bfOC*f3wrfifYqV4OfBtTM-q6-okV2y8{E~rdGn`yp57>Kz0@&k1@840i(B;(`#i3czHx+LPlW}QH zt{Hl>`GbXTOIz?V3yej63j}V&zrVdalxI$x*oqJkBX#Lzs{Z1Bf043)JZ6x2?k)80 zR81Zlb_B0tk~j$$PW*YfMYPEaNaJfAFpQ(BPmYWLIdx;QmL0Tmc*XnH_c^l2^XT5F4?(dvZYZlY#QG zm`6Ad@$+lYa<|`XUY@Prf@5q-j>G$i@-RYPpI-J7M zec4_0Y`@y6+3}WL#HWD{x&ZLoumvMxIgAL8^bkFUx3YH{f8~@TUjD>t2YMLh0s#k8 z7@c5A3N&{@#gp|Lrw;dUPCk(DwJ%P?Qy~G!oTVEWRA5XBP$?Pk^*nC&(Zz~V_6A`Y z{cplX8G9cOfOPvJ0tKoFAnHSxT7AKu`pr{%EUbF>VdN{Uc|D>jLA&_VlOiCDZ{;@Y2$K?#r|}%YFWO zvqGET6R~RQ%4JV#9Z(*!C`4wc`56wD~i4lgA!v6m~etO}+AMm7-f9c9uSq!iXb zXEbUvqgi|;vk~=|1jgRT8but8%Ejk0A+RtNOcboef9>)4|gFvGydil#v+D~l%OYwn~ zhhSy$n2Ypfo+LspUDJyO`xFb4%XqB6d03qKDff#AdKg_)4b>oO@ilzsI&V+Zi|VhQ zAeEj6AFl5R1%k^Q40>Z=pO1hO`x2&SMLkQRe{xqiHoFHty@3S~#ESTCy&4|LENXL@ z*7g=$4G&+f-q$X=wbJEGb+n2Vvbyd?_p#3`7W*fN@s(tgA9Bc7eez5y-}jHKw2o_u`M z7Jt~<|F)2V(Zte56QXV>zvy!Jc0!dg>sB2K9TSU>#>Tg_+W>79+P?uGe!qO1fBWM8 z4)~tCs(<7hyoRmOUtJkrNpf`;oYPjF-iBu zvJATtddkGmwa!*KjVd9EZl>etd(?ms1<2oe1&5N9=%v1%iZEX+#_P{QcClw_ zD@}ov;_E9I2cln&Agi}kD;Y(HYEk?`2_OVZ>id(ZfT?==MQ`iAc2@&qA|F+NN5%VNwsw5U)-qOcCJx`|BJkaI;`O^1fB&pl26XSo zokXfx<~))2Th6k<*@Lse;NtIOZh*e{~xj2DffpAs5iL~b;ebEMpE$%x;EBoUX z*kKzO@rd^U1}8LJ^)`RLfBH#OP=qU8G{)tTsle(8Va%&tr4o0WLeqXcc8pRAjs9tj zqPk$6zvQXC_Im(lD2e1NdFyAOc<&1nCOu>V88S=ucGQCEL=j&I+vfW>KxPU0+@URk z&T?`nkXRmVc!K!FHTVLNHu+Dc2ZgDc<)frd0DydG0v*t6`(zr7f3x9T0;5}8XyrG0 zfiBU*l!FNcWFX(H*jhg)$n|yXwJer#CNsR_t?{Hs3gv4 zZAFgHFj=yK8CogQqxCCyQk`d*&7fA3`_1W)V9$ zEfHLg*Qfq5e^(yclx(bSf){Y3H160~Wc}G$o|D2Y=gmfjZd65x>Z}U$lCxy?7Zwan zrpI%$AVfIMY}D@MKTwmlSHFIQ+kM?br;{_5&_S72?j#dp7K030uJzz((Ym+VXBVPM z$qg(uSDCI>Bnb}P+N2En=6}6i=}jOqVTfaDK-O}Ce-GWI6$q=nm2-(ZwHK6_p{5a# zH>K^lBhrbTOy$nH`O`T1dpP2UTap@pEN`Yp!Lf;QS78q$fW7e+^Cq-x0LBN{ z*MOv`*p!>TWO+~BVnKuFk8djVk<}D2FY{)VH|VZ+{js*_Nv7DHZRhulM92SUj1@ki za(0~s5R*K+jz^~q>WxugR1R?6tHz}+$~;Fle+I7G+Oys}F;$h<7tOZ&fE%GIE-1+v zZSIuLqxAEB$xFLgjNK(d7KMMt#4?3^5czNUkR%VwCbp>9E7S34;J5`JRzsm@@YY~M zXT{*NeSMhDI*aT$h3lvbn@lZpIg^`PF^4hU;ZKDHfcdB14Xw>)dl@sW&ekg{($rde ze}IIfqg}!O#Jv%y zA{w!u5t76~pt#T(YN+auA^NVJXt(k=`>s}=?|zj=T6f?>(G0m9kB;^r zi9O~oA>&Y_&!<@@=&X3#HA3)trHX!ye}=lhMZDiC(tD8AK+1TYtu8 zFMDdu+J$1r-V{2*Lz%m*$tmkgsf^RzUd=9UsMfySb_2y_M-sm@R6>%reJqs(F0CUk zC(1^>mrzBAF}wv#CSZ0(n{UvFF>A*e=@4DM$3S&WTCgxKLNQ<_cJ%Qq^GVY6e~Sjc z(XG17nUD(xk1J$_8-RO?x%GrSn8K&(5#b@Bsu*t~FuJ17>I(7@ylAJCEVA?^n^O(F zP!Js8MH>bxvDEH8eg1wD1BzN5&|N#}Xp!~V46;Q_bg2q8QT11F8Ra!d^2faYPRoJj z4#o-5^>Rinz~RC+B)CyodF5W1f7?mXUU76|nT1v}_pOZ08{d++7I@P zR`{Y#BWZimYTw5Xz$-xNulTloE|(kiLpn6ME_A8P-u)*qnq~lJn#Sefe`V{q@!2&7 zF|fsP4y0?3z68UK&#}I%o%d?nBTm!PVhy6&ha{q&J~Bwf1B3`2A>d@c#a2MNxLGp2)guUL;Z^Y-Y8-oKeG#TL;|u@ zqYN}^8)9y8(9{m;U-V2bUoDsuI(gZ0o*Xv_HX#B#E(I>PRncC9kkJ-W&zg8|l0Z?b z^bP?zV_nO?SV z7O=_Xxwd)x?PQ7wjQ05N@+sV8Kr9FcOQS0qPm#@ajW_fXB?_5i=LJt$&p&o!%QZ!3 zxc!7@;bPztwi|3O%y5K|=S>^$1|vXwx>L7vK)ZfG}>%`NB;Nbz$gU-BE}5rTQHQyM#ujHEmSoj8|V zLBr0qy1P7GWqaDD`tHlGhA8;)R8W7wzlpyw) z2nupN#S_yNAj|5DkhreRtkLPQkZPmPTS2u^HtP`f6UqyeiP6R+t&h{6zuKJ1#UUw_ zQWRY4`NfB_+V%#O%fsd}o54?$7#PPwYu154$9gs zyoK=$lF4kOtHhpf-L{n#7m~W!1B;31D@xVvehAT)l0!=al7p+rG_mSJ>iYD7u`D3X z2E*sns|4Ph!G3SqCZmVA?F1h{MkF+H4#2f11Y=k+gj0j%(-2m5|)7>aBaj;(wcJ zUwtH8U9|P7VFln6HRU=8H&b9>Ki@;t>Ma+2&+puolWE=!(D-c1T&XTqL*LPB|lw(%XOYC57unh4o_H#78e?NVao0~~tl0d4yG548# ze+aao?ca9~#?2Dw!q<`*G-iM)7&?z(*}Z$~B&K%decmnY{YeJnowQCwYeRE{`eXfu z_z|A1&-}zESKsX2gz73`sfJb3@SRECDf4! z6wF`+_~aj0RHpDrnTc{vj0i?Ii7iSOIMd6U`nvB}zAa|H+~QzLnys^Y{;?; zWGdY>yXBEx#v2bo(3S*#;PWf@?YR3LJw9jF_RLM!hEc4UjF%38 zE=`c3o}4ToBxs(<#2HX)Luan}e|IJWrY877;~yd5 zgA+ef7So=pGzN_+DY40rt&aBF3;cxn7*XT#Ok3S{pzaB7AVq9nz3c z5kK(n!CKuvK97ljRQmA>#nKl{i_oF(L3fr3=IZFMbE0}u4xl{-*A+irIgB~)Uq_$O zpxYrzRwAL1ef>DbOhq!oe@ZKp)f34HixW=I(@~UvkHu}>IC5TUBVs7`isvxvD;WJg zEGH%DVqM{4$t+nR)S6NmC8Eei{U#MdV3Rrpir2S$a`kA8*MB9&d+6p0n?>JVkZSWkW4daUmSFli6}ZR6p~Hf-~xR?$MAV#+``=#!jev1@V0^s5tWb9ixN>6x4+y_ zdAo(+^&Qv$1OrO@>!xSMl5_@tA5ES9|R2W$-W!16S#e!Q6 zNf~%j4!T(cR9zu~PgA^U9m(2L9q=f4^2VCr5f1`uN9``%t%n`GdDSR# zUnUZ)?@hWa!9d8UQSjuSkFUld1G-90lp`Z1_Wt7=|NoAse+LW~L3L%KF_%3{BmTO> ze2;nRF#$aNbbRh`|MQRt{U%&(Ud_*jdkFpEoa;?ZFu}ns#1!g;fkgpi38XwaLuM!J zPVtMYf~20-vJxo$S=(7yFe#8uruyQEV-L!SXMdG*cxe z98c0eX5y)@e_==%TpWCV)=l!HJhxSBrgrUygA4JDlH-4Aw3)Fg6PO9sl_)9v6^p54 zo224H$Gq`ngRgBmyL`7kd{M(g&A|3O*VRPgX%HL09lUKOf91z-Q;sKg2Vu`PCy>(j z6hH!1JTDI|esTP0U!OQTbdc^f+_w!z!%-(v;)LD>e~pnR?#OqTd`7*g1v^AY=Y_EP zRt!1kR%)}wg>2q-1qq6&xG`G&Bufi@Mh7fGWWPObtE~jTVx-=|EGxYA=IWLuguenD zLE)3Y7zT`|i8VL_;NRs`KdsSz4H;N_veHf4y|At2--B{oiB;@|Z&KyIuE+|#tEDJu z=c7@wf1k52uE}*ZYTw`ajlIesP<>W4bf%p2 zR4=yLn7e_5yDJO-%AZx}YLOquQktA{z*8qMW)Fzn{Gg7^;EyM9`3gh`Gm?woovGP! zcqXUXZ$=0enNQxy!v>G;3tB|Ne0v zM#_0RY_`&)S^G1RIdzCF+BNN?w&fV0zg_;yqlx-2ZQ`n2Zb}F(YJQMW6EaumFc99R ze_8(9sIfk*;jwZnKLu@*2j9FEIpNN)-LVgh)oamf2HL%h>PAysv9bTIDO(j%C?BjmCPJZjJK=`pWC55o0_h z?|hS5?-EoNzVVz_$neCCpy#SapUQgYc4=-fAng;eiR$1+5q(<<${f2+-s2s0LFbQ2 z&cDFAcjq@;bEotJoLJZ zVc){#*;mP-cBsR*UPlE2evSIZjwu~9ola&sG0tL6yZ9X}mVXU)gygX)4ZmYJhs7>v zIlJuEQk$T)GS}lDPfs~fcca;7kbjm7X@gL8zr5fq&+G=oCEah$Yk%9;_|(Wk zH-QH9{BHh^9|u3H1A_P8e;e=oMJ6T@qEU;R0)Oa{e2@}FJT!U>k2RZbd#%qYE&)!F zTjATF{9t;_VdH|qfjx1A*Br0tI+kkWRCND+g*HGAxmtw8Z!BF%}3BqI)_&; zXCUs=JkNzOh~-QFf6R9Ks#3WXPm?-b5d3^#pZ(yXV|pa?r332sXxht`q=-S$5T!U( z&myG>*olRyw<=%w0JM+u$%G0uijZpkZgTe~IdjFG_M+72Mo zbWE`2-=d%m%6pVgYQ^7FzYEv?BVh7}z9H%@C|EK!ks2v$d|Jsp$d_^nhwvpZq{glG z%}^VYsBFZ_f4acdIrlf@yd1gBI}jZUCyz-5S)^y>Kb7iRWi_;Nl$PfJkp1H1q zszfiwTm%!ryVP1|i7zL8kvU}3ZM9wcMt3MjI7R&QYMM3`;6->N#lxq)X`EltOBeDc zn!R|wp9GB}L8Xa9jbbdXi`gyI@iry;3&GfSlln)$e^E1MkfG?!*nV%@%1fY)5+Ph3 zS?Fq)A(ylNu2=L8~ zO8O7ldTSAo6V6nlH4+D^sq#>+N~m?q z5dLs`e?MKFQErNIeNq+12oPz{7@3kPfd*3M-WS)I84J}lRy|ql3scGWm>g-BiqGGb zC)F_ZGwe?XAa{{az4k*L2=V`EmphJ85^a0N%{eNgDSQ9QTgC^a+kNH002dVQbJ(V3 z13$}XPa()%kDpeMy+^&US3bwuZ|zIF$MdV6y~aR_&f*z zwIy!1eEOleflLekr$XGS7Jz)xpPwh4EYWy5G#)<8w-P6uVD9M4&OjasY4lADuGWCw zu2ppZd13`Cgy!LdE2Lq6frI?{(b`F@U9(y3q7QW}MkK_Oi)N9)9XR3Zkps ze~~)$p?;Iu+GFqN`8*{=7Ehjxd7pVtF5jm^=iSkMrJ_f{@RVH2$A`zsR--qrpH1)B zS#P@2U*$D+(as`e+@E{p58WO-d^u3`UdNVQ=mbfKwQH^^I?aM$YE}j}L`FHdWXbE| z!@r?r&{E7~H~`+dK^-QFC?av9rzKB`e+?2KMFvElB&RCdXc&7~pmnu_#yxe|xU7EPiA)>7)u@h?3~V1CZ3QhqxJX8(bW@PG0Zc zx--C`<7qm3H}&1^AQbaGi6QJHR}llT&Hw3|#=VLEB*kBLS+gE5YY+b?p%1)z_cPGa zDW12Z#=6}-&TMF_1jRbn%QM#ZOq*X*y`!Y`& zRsLfM+9aU7oQVBuZJ|!6S?fwBhs_5agJ|$r5;^<+t1$T2YuTa8UD;>Ac0)Xlg>`p8 z+Tk!gW$|{`juh2fDE-*ZXu*?tFo%VLVg%-IFSf6nv3hQ|&DNp-gz1IfJf;0j+Fwz& zhf?7(kQGc*G8QVaR;;Z;Tq)x#0Lf91;~`gXv#;@t(efhBo>VY4#ZYmYFPQ#o*$7dKJn~N+Fr(U_W26 zCz=j2IMN|!C|3~u9G0-8Y>Gk!Fm^}9j>l)B>As4>`^PnQf9FT!jHfT)z+mtqVaMt5 z{LyPKE>Oj8b5T{}C+8WF^Wh(?4XdkNX#wG+V4tqsC#X%Co+uo=30qS}Gmv z>MK?YTz9V{li={*%UVMKVL+b02LO$&nz+TQ>c{q;b|9StXXoyf425%fp4^vMY>i&W zy$)zc;&%|wj*ed&f8n`64FcF3nQCX^p}S8?#4g)0-9tOn^nX(h?OfzVt=3AM8jjzu zgBm6(ixmQ|jl26q<=qC}UdGLnTt4wcKGph4?dSo22S~qk4VC~ke|w!R0`qA~WzAi3 z4u?Vbh?uFRVJr~UHjysj1c`UsxvkXEu|V`=fzZS>M`#)&S;eewUu+~QkZHlV2$ien zHzgN&vrAdq^M5eU4=4{rcvfn=^x8LPgq!L0#tea(T)8gAsjEynuelD$K0%{imswr3o)ho{&tYI2^l*XZrx>d^Yx*UVY2j=z~82Y(FR`2n6<6#7Jp<}-8r+Ha2j zsGN3`O_WfpncV&gs zMsAL%hO10#rFyW@VU6UD;mOh<+E^Emz@o-yUK5lyA%8=iq~ahi;Y`M5=n1sz2ch!L zusQoJf&XZlUT8Q}IG#nrOPtdP19i>u7Wq*T&H@@{+>^mMAfBz1g*AHdIh+6H7 z>9fcTfTU!nq)<0^Q8e%$Yd+dtYt={9gHTOJFWr-3kuab8|PxTF_&)S)NrRTA! zk$*BzC918&Q((pjdg4odtQVd{8WIqyZRDqP?NeZS!q=OUdVi8~>yb=3!j+YRMah`c zCkj4Do?MqQSc@-3y6Yz_^&cK`-maedBRoJg-|4tb4?zXAzUF70yj99(f2MC5;8b33NK7$ZfA68G9WTGF*q=n(Tpw@12i@> zm*F)5Cx5iNWl&vP)-_CUcMFz-ySux)yBy%)?jGEN1`omAgS)$g;10pv-TBDvxBK4i zdj5Vz!C_1tYp%KG+PezKiIr68g-z{^K@#?MF7(U{OgsQNc@=wGpdB*{y@ngWbH0rDUh zpqi%xh#5cy{0k`AJG;;u1D)U1AUktQI}r7|i>SSWr<0|*h0C89tn~DMBK`3eVF1Vg zO|0$RovkeaKs!@_41+uaK*8SqU1$lQvbO^mgDikHW&nFLfEq{(psp&eq6&~yQB+q_ zrGI9453K6y;9&3c|G0>%s;NuT0mOtA)WiWG4LX3Nx~kfrpK2hxcmC#d00p&o{hvJV zhJW1U#npt>w3Wn}8UL&SzzlE$IXPSYN&7$CDBhU?{-*ZsYUX5b`&R-0m4%Co0}msk zySqDsxvR4agT0eEgM-aq{M0NgodNFlPJh;b_fIE~4d}1JxZ0V%m(#@p^sfMawgez& zX#%oy2K{l8u>Y6R_PvzvPVaJ;{}6jGgv+0pHvcjQIDZ_K0*JagIsM5Y|8JGkf71M$x`_Sz%Jgi!eSd-O z|GQ&AJ6C6~f7Ir`mu+Hi=WOZh;{30OAb^>r4d@Sh=RbR9Y4?{*URXg|LR?jiUhci| z?C9m~->YNC;Ns!(m-U}`!eVkf04^pD05dlmfa$$S#qCT*?QLz}Sv$l3DW90-d!1bD zoje);r>m{)?A`6W|5s{eX=iHor+@0Et`3aqc9xE=AZfAxXZtRK|667base;_K#l;A zhlvH_pHzQ!${#WFAMtw$e7zm)9ROxP8)uNOr5Wh`1K!&i=mr9~IJts+z5l8BZv@ZG z2{5%Zad~g@_k#leS9EDRGkXB{zr^oM{;m3d+fVh^A*Fslp{DkBHl6@ekbfCGqk_H5 zdoxh||8vIvCzph)jg11(7DV;m1^r(upsl5i=YQM%kBKJe4^^uF!`aeV!qNj|s$}V6 zV)3u4{!1?H0(@_JVLNjh(EE=3B~t%$j%?l=_I>DB{+U?-dS*`c|IxkoqKUN~$k`dd z!SR<0^j?Gi;rm|rKWqVvs(&IfvI?5C|I;pi8HwAO*qd6~nFCnZIRHQ>C!i-h(|ZrF zu(JcancrL46y)(&lK>bQ?Cf3MT>uWQF1`RWdnfom8_B^AU=;o%`WNB=FpB(*I01~J zeTkpfV3htFy+@Jz8@)%7{~NtWQTQ8i0~i(mgMYY~0E|k1qxTFd zf1~#ds(&Lk0HfO9=zR_9f1~$&nt!AB4BG!e-0wQz-{@Uu@*l*^^q$`IACMWq2>K6X zeb4`1MjN2*KP=uEn*9w~-ou$${zLulV*U^K&e`HWknNqbg{OlB$nGB&@3MdTzn9VK zAMl;K^*`WyA#MHv-+z<;gBtVuTK*B4`F$<+|A6nr9sb_fcb@N4#rdyJ{oCf9uG2r@ zJ2&Tlz;|vg|A6n@T>k;zxw-uVzH@W`N1fiedHe&ubMyQMeCOu%7yR$Tp!DY!_-n{9 z{e3$A-}U$xR&{Z*w+3lin!Z08{KG^Z=;CDQq096>p_t$0?|)za{!RbC5|IBrT>s%M zB4Y31P0#k;(ey0ut;Ed!-p$O+ten38B+EiI+3+N~BhmR~SJesA!pdOemzu;fmK|*~Xla;Go#5R@Q65m1u zP#--M=9n1 znKsG|#U7P`cIfUz)h4_mjVJxISaF`x=X>7d(=m=8Kz||TSBYiX-CfLh)<9s0RZ*iv zrU@fIzMiC(wYbox<@%ZTak)NJCri#+fwK~xGj1yCXy`fr1Wi@F(zpxoEB6%Xa zB7gVvY~BX<=c#QZ1E8ulP2TlMaavokJvIAN*HLt!Ip(r>UIJY_S6qxpDs&!=OP?ur zGWi80*Bz|)WoTE#A)kd=#kDM^x7+k(a!d;3P*qkS=?Z2T&lK7CC$kLO-(_>T)=QhU zSZgIj;`B??(Dng`5?XXDav+Rk8CWUsTz>;%RJeTwC&oA6pk+V8G<8#?uGSMBw^J^6 zo+f&>{YAm;koCUf7HQD#WU_@I-Dc7dvS5;UDr+gdTp_d`t(L6Zcmx7{uJsej|-b+>vOiF$3>JyGE(Wg}hopTDVAIx74S$}@a zVApsjK@LT9KQrZ7UDmkcTjLZbN#g+{pv%#4F}F^?%oTO3UZnaf}r4OC4a9eG*W%Qo{C6%sv+r7 z_HdZag!)4Dtt{q}-8h=J4!_chA zeB;`19&?@)6RM5#rCeJ7QFqQYa?ZVaMFVWnEij3FpR4F^;w8?m-rY9+P>>MJLWWhz6fyq!a6cp0udFRfqYS6k_Z^p~j3bX>t*VC4<%ZdW zHj00gcdt}5dIUKev3?`(is41vvmir3jyiZt$dtu10h!5^S-RnSeSh~(YN!5@o`8T9 zc-fnk-qoz2(HA{RvT#0PW5T-|qwRjgAC4*@!ZeWmrS1uFt!PGq1C?H~EUuX`pF}iK zpmxvCrZ@VemA5Vr0vr( z#qw6f_+-Lnp7%P9VtYMf}TF$he6OD%q}WYpBkfIUutwX z#4ETxr>;wJ2FONgbwiT*Y)>d?PV^5*Jg)E=GzAyTk{83l!~|J>S?UfL%{xe#TIvK@ zfD;dJ`*FO^dQeV`_vwwtvuyzG^-z6wb{#&*{^W-f_Qo~yPJhw8cfRf@0e^ocG#C~A z_6x&F$MY<%Lh%QT;DyMqDv`L$j_2PRXws5!4H=aE!2GPZ{$*52R|WxEwc+@$ks>rYFwFF&*^qc(}k$)5u!ce6flphb6Wu?I|1=jcx zn&QWgsEmSw9_8#!+o_K2gR+w56&TAGm5|(RJI%GYdN#ej5x6dT*w-_9V`bw2t1P0@ z?8t$x*0EoCwqyg_UUslbOZ44f)1z#oI;BjQlAgKU(L2Eq6sc4$`WYtG*@WUk*X`%p zxWODWtA7RBb{6)JP@{6UBzg^gmCe_%fgUl=wAui}W;8-M9`(bQ@V;m7Zw- zz+vRaj(&Ux-^W`$3d;{~#Lq@@f-+%HC1fxxx3LID)sqwrwOMQdD%w0)A zM9^n9^+K0AW|g&}ngfpQxcq7Jxj@WXsTSqmVR%8~F!G!3%%ms z&Jr{dGpE|3ieU`{)3Y)W6gWNzZJtH$_y--^=IW%$qvh{ml3*{nAv|iKnX`@y)UGY1kI<3qXEBEZILayOqZI zgH>Zjb>U+f^6{^|CEFv?$gr0_JX2rKyd4`^orW6mJs1^MX%YkLm`1JRlYJcCu)0!A z$Nop9Gr)1}7A(rZLT!>7y@7;6#Yfdnqkm!|%u^O7|Lf&DuaBN$8!P9|pIX$QBfQ!(V_dUx{G}30$wfch*Xch43A} zVMg@x9c>iG>z06la!Qn9EjYG*vhar9b)h;q>0c%-Rl3-5916hiE{ehC8yH%ne}BdI zx3tmYoe^O0^&0M%M_QSz+NKi`*T)%kdwOd$wesq)t{jkN-|B_mGvn-y7|liJfPo8` zlj+0lCw$HMCQ`MV%N(^5-p^0t(m2950Y zY{vOrcivLI<%hqp$=e^mm;0b;_md~!*>y1Y+7-7JYD7+MCCPCDPHZ-%*MC;>%d@Oh zmjEdBwtYdZd8Y;JVCg)OgE~X^q7I`}==;CF+nt|d4dIFoqnt<|1`Mp4SU-` zvT}L2LCvHD`%+v|8g%U97nQOA})DH8L;y-4V^R0(XJmt z&fG|d+i?|mxIEFiu#AUgmqBI>Krg`wgRk?}cGVl7T@H<;js5V!sd<#n?e$sQt;i%e z&gxcq@$SSmE*_fUgeEiwOhMQ;z+`*N+w+@d6If!x*VMS9+VgG;$$xZ{lkd#WU7vo= z;?KFD$gYuz_n+PQj(roap8a_vAhK(t58+1tJ$0!C&s)&Eyc~|^JH(bDnzPAnRn(*s z)vQqXO&TjZsKHh5^{0-n)zcQcH)R}Kl+=x>kE}0h4mIbYHc4v88so)7=_Vs$#H0p6 zO(@soluXH$R_R+pvwwz6LjfB~t=wm6GEVPW>xxpnp-T!9I zWF6^S!cfC#d(3R3m1=LUg<@lAg^o<>e_?1QO!8B~vV;WeF-0+BnCB%9&`#A5vbi%I z<#p!Mu<-3ch)}2YH5b-2Kdv*y=ev?6(vR?GsOh0*9DmxC@GB+Fd-O3qiQdntuv*a8 z+843kR>vYnq}Z1qLI{~}6pX-&=tK!=$jo>O zOKjUU*Bud^al2J4eXQ}YjTy&s`z07kXaV|DPqu1))3e&(ymOF-nzrA?Y-jDp4`E1R z5s)S&V1Ll)K2v_ksHK+=@a2H-+iy?daHUef5sq1mis<;Qwba0>C`Fx8n~aATtY?}Eyn zPGDIlHbiAPY!xY-$M>y8E<&n>AB+m~B5fAhhkud_>bKxYa;zGwlGA$aTi@qG?#auR z*m9*2XYoNWIlwTT78jb3XPn!r`FvY{;cvmoUbF_mym1sLb_aAl0qdp%q*KlZV^1iH&7X= z{S?(a^y^qQ68d_;%zVcZ4QWA8rtB$C%z`4uc5-62SR19#ek;Q0n`8-@=}=t*8+pZT z(7Dg^W}U@XMV1?SI)<>^-^Y;9dUx(C)dn+Ih*hgD!FLW8g-7 zP5>WIJ4aO*mn>`Tvr5!R*k=mG?-M1#G>NF$5&CyJJ&#rXKdoFS1yvx&reM3_PF>GP z`L6@yEOv0(HR>dostY^wxn7FgX&{w{Jd`qds0IXLxaMfSx~jTQQc6H}0O5rUReu|S zj$!@C>l5>dpJ0X202UzN*QkBEw$V>i7S$E1u&~vWcBuzMuY19KiHgK!vbJ5FiAV1C zsjW+a-uE-$&`nhtdO^bq4j)NM$HD{EyaqltV)2Y}Vo9eQ=y!9-mchNj2r(mv=0zDn z$hQ3-<8nZVoz-jHe029~EcjHuqJQdQVjl~GqAoC_^thSt^BW^Jjs$cWr33Xhg-RF9 z=(~=olMkU!`2GS`;tg7MTXgCDi8whn?QGsHdnXSed3f5B)7fZ;kVkjRocIziAiQs{ zIS#4YkY%4KlVYn%wPRBFSM`FLh6`Sv9!ORw{2^=GapK7_lAv;t-Ktf zb-7E>y!keuJEGDMcReLp;D41GrwB!(bQV0>XY&d-=%hAbxo+~-`)uy`P00{W0AW9g zCXcQ_aN}RF$~_D~Z$eXXW>i%ST*(A&Quvq#7tS{rqZ$OzDRfT2nK6FK2o}B< zHMY%Yt;CC)ggmNfUVnmPL30+Pg8CI3Yn^QgM6d{?flSt2iO3^ZEoRBK8G+}<9^WnU z%2Ym_TFf6FsLLEJke}>xtE|rqlU~SK(XCb>9s$Xxz-gV1JxJCYi7t7ZrJ>=MwVpKb zDvi?Hi9y)rt;tx|&Z4na9wIF2Yjg=XIDP%>i7XvpPQ;I%BY)0NK=7?GxQFkh${tci zkEYb-7tg7ePCf>?6(=%y+tDrNvrbY&%PZD%Vw{+=n1Hr=%d(ko^@kv%6Iv{8e+(FU z>t8AEDm*^nfn=)6T8+h=VvAI-LU;Qf`mdMd%Re@4KI8~N-H_#5rD!gSvOkn&iFC6f z%)HS)@cc3WeSdPj==)5O){azc?_+?wzJ$8Ggo&nD!Rrc>bGe%sab`4pqtn-$GS9f0 zs4xw2;tafmrw_mLsANk{M*o`v_8vsyQc<-9!BoNt;9Bk*;!xRz^6-gL<`SHWt z$R_BLz<-{VZjxv9OTu6h(7)K#-olcSzdH?= zDDl_@!8iknwtvM0ANeKym@>*@%`>c3eL_!ONPpCh<1BU=&HFpK-{hCOc{Oy?V8!Et z@fCH(&n99B2`*~L8LLVBFCRvZ3V`QSKpeFP>6W6wiu`zgS4Z8He3M@SW>xg>w@IwU zBpitTtBCfv@i#d%Ie(UttCCj^qY@((_hHKv{feC^yik7C2H&?eMkx>la*dY8k7qZE z8Gi5^dG&dKo)ADJg7}Bqopd%+*Q<72i${AoOPkr;Sr2sf~v?w z{^)dcdCxsrkyc~Hv?S^G3K!Sd8YWNi`%P*eaRQ!y;ehWuWt{(81wMJ3bB3hctzqQOGybGHtEd;9h~YIw>EU(VL9R4s?{>ShJ^3|%ezW0TK>OGW@0#Jj)qm52Ng^}^ zJe{0wjaJ4T=0#uA$&;TD@pDL~{Qh^A)2^Rg;OdoH%#8t>`qh2n z0r$?~zGYXGAnakgBEw;DD#WoK4=hgd?*(~BBzSoagkJ1f`u*o>EvuQG@9-)S8GWaw z<15wm`P9|H^X(8aw=g-n7~7wE^71Z1LEYz`2P65pgty7cYK z8jAe5@mPSD6CNz8LteRW#_iADH0?z)PK)OXAJjlMj>$oQOOfAv0Ue!|i}_r$(p(~R^xE0LYKUK2^4Y>7c_o36e3+jgtFRnk!iAD+s(-Cr&!4yLT6mw@PT+Jjof0GOKn}c?u~H}O7s2vV$8vgF6+F|j zV7|Ub;c73>RdAZRH-8_W6I@wYD%=uWU74!9rd2eUOmYevO>b>>32|P!I2@ISgeLLt z5>mRocky8%*6_j)YT*xfBoZjX`JV)ut2O&hU%ln%Dy4B{gF?_HwQ|Q=%pmvgXfK&< zzd8o{M7t|}k`GSg>j~#eUHXDX<60KyqZsekc2RjSRd`HGbAK@}%>2AoQPJu24ctZb z7*MiMIT2Cie%`;3OcKd+Ut&VXLDzt!c(@X5=YN3k0ZWy>gV_G&_h$|T%SCx*cycAm z1p>J=<)y_p!M5BPFPYCtd5dt|RadSWs)v)D@~{2N7a0mEUNHG2VY$mQG#(qW zT%1Ju28>Y!n|~8eXTRUpd!r&KfNMF|Rm-ivT?de;V2>)`3WXO1wxTKJ*4CKK?KJm5 z3*uf`w(do$Ec!kjX?7Ov`*eNP}zOJpY zqipTB#koi$YAz*AfcpYJAJO^ym`vT+B+MN{oH5*bQ-9^GQcog!pDyiTmS0f%dh(l9 zxb8MCAxRbX32PvCQi<#h?3QUHq*kuda>h?@;Dy$0a45+VdNH0zurRmTN* zxpcnruzzGS8;LkuId>;70AoYS1V0dYI6-x87X6{u5Ef<~Lh%^nolwlZzek*o5U6#Y<3h?{N-?UQm2y>jj ze{dvhr}_}t;*<-cFCnZ|8&dfjN$3n^Re}3U!&uDpiNbzJU;a>9WOjvd;KE8iv9t;Q z^^b5tPb;k#@H7%^`C)i=)1OmN$WqIX{1!Anh&Nto#CZB`tJX&_Wt!LNh<&|_(cBZC zkblk>cTD>fhFL@u2VqurVKog&XD2%M+l~Za=1w0U(>WD5(rg5ZQa6+oHbG0`^xmdj zcnm0GQkAu!(sfngV34iA$Y|q^1Ag_P5cS}Y_k?f1gl5^LsE3GKaKq&+_*BETWRbEl zBleZysY;`CJN309@@h|>K+3dLQ_=5CK7UjrFu{RAO-;M(ZHeAHlRo*79~QB<$nKv% zZjIG~%kBHe?Tjs*62Egzka3mujVpXBsg^3eNDVw<-Xcx(s z=W>1uHikk=xETg&YO~EuOMA5xi9$?YvF`+GOmm{CTGs!X0zLM{B>S1cm`6U#_J14W zq&(>wW5akoNKJkEC_(Vy$h?r;a8NlQp8?dn@k_<>2Qu@MWZj)l^RWWqh@?2dP<+N4 zu9sF8+J35K2Mc}>b#pYkv;EXqD>6rAgDD>s53u2Ys3Lqt1;d4ymSBqeA|aD%Oj)f% zyTIG7Y`>N^)M&~z&4CStjX~-;I)A4HcZYQ!nkFm$L&YR{9&u9BXzFnFSN0dAah_oK zrWryrEVWa1p!0#j1S(WX*38UrF_L&lkZD#hZG%jHRNzu}CFamBRguxtLd%1H8q<6@ zThUBAzvxVj5x;hy_?CurGnu(DI1 z+y5TH<^bp86<(!q_WrgHAzn%4J%-j z9AOgW_k)sTDLO)o4JVv#L*S^S8@vxp{M*;7xsA@CjaK0UNLbyx$}LZd@4hY-%zCAk z*UQZEJ$ADV;3}(F4DdK&6_BQ?)EG(K!YX>C!e5u$*HU#&{VI=?(tknd=Z}F{IvcYF z@kLpgcrW9up|Z@2O^X}UH0Y9P(x0-F6ELL zCC(`fc8ggBF!%Ed&Mep#?$%-MU>y%@$297g)G~CvT<{;+x4#ZCBl`VBc`#)Mxza*L z6?Ir3@E!&$@VP_JWPjh3RfarokCI!}JC!wG;9xbGEOpl2!d=@0`YhR1m6J!PABpWe z{mpdP16c==B5W$BNnbl#tYtu1nBz&ziy=O|*oCFIqX$>1r!%}g*l)aC(IiU)kEyjy z=qBL$B*2m|$S=|5j@9m#rPaK*Y>*eJ2tn=X`E?*tJbxTxuzv&VkUV9^nb83!(V9%H z;U+VU{)?gJ7!suq!h%{-^8^HT=kSs04Wo<4K!7=Ian~Hq}*4}^N#V-tRtO3 zL6oe#O^C?uhwSY)4GqU%mF0an0v}AW8|IVdB*8-zmDYA{?b>zk>i2w=Vpk?wYU$Kn zSs5QW;Z|P9<9|)1?J~}M$2@my2BtVL!>7(BAUAMHY4rH}**7Us!_#zkyLXWD#Vr(N zC>=q{GyBKQv!?xB!79_)C0UIP_i+f-d|piihJh`WQoAQ&=-w*? zW=x@Wr}fq%&}(u$b2`&WWJocWRJSqe@Pq%DWu(vS3n^-FXz~1(k}I44F`@S->GVC< zu#@SKAVJuYt9_vK_4W7h+iWJeKHkq7Ze1fdhksdfABwY;d#p=7ay}CH>exw079rNs zf6L7E1d<1`P>%*Eh-IYKW`Uu6r5%r=_EUggd+Dbx{Xo!V0Z~izP^?>p_7Hup+Cit> zVEXIV)ffU4iAc99KdIYqft|jU&*xbsN!CMYHu-)M=&Jy2c8S79l9kwWqRZ);)oVi1yD#+gI=Ul;{Kp)e) z^~8zr(2G`w#1Z#0Mby%kTi0Lh<>T?g3b)3!SPl(IBo)pWF9yF7GgGX3a!(~fHN zUTNSHK5^pE359>pMTG91s<{fxZ)(gPA-|AWC6R(ishfemD5JAPzl}K1v2d&~H1X;cy&B*>+?# zPQ_43a&IH=P=!PEUR@>>84Qn?OH!~XfS3=rGI9K zNc~KiR7+lm9riVJ%ziOvc^3Kyjn?;SSz&tpIKw0XZlrm31sWYIl>|-ETfG^^?a%Q~ zja@fJy^=m=EFG^rKrsK5MvqBv_?!W!O>o3Sdz2W6JOl5U4~*EI%A1C3Yewg)47|>8 z##;qNp+f~q$B1SKbd;$8DakZu4eo70|j0YKGfwdb)&tU^IHEzE~{3-MSn|O{8NC% zm+UNMG%8AlTA3%83vRWhz!@-Y-@^Lf-3MV9`Dxvlw1z_RP(u1Tv}7QKZn}pMhN73M zYy=^w3_0wpH(EXVan85K18(7Bf3&Y}hG$=Hqk(2g<&(1PiQ%-;pX_-=?0)RdB?7Oa zal@i=p32Gx2la@co%d6gaevh=*n1i>!u2cff~q34d2!8=%dmC3QDtsDNh;_vdqSxL z#td#|zgq`=r-U8VQNKQke<~>l>*~>GYrQ}SU2}s-XyeIYNv$30xV?^%=fs~Ounoj$|?t4T}YyhDwIFT>}Hyd zC+DfJKXTay;bCwfv&)-uz9oW@^P@mXt)7qe}dH2o>Vm=MSn*I0(o?iVGk zrq=|}7H%~+2NXOsgMa+lT(ARpkU~Vk0UiLCGa#sy|@7Z)!|&a z#d~rFzMouoMINIx&@9^5JVPL31@RkfbKx^s5K@hEvG2Gc&sA5TZ` z2w`zKj>w}lFt)j@TtrOZwY?Z7Ec?7tPknTaO`9cGmXMvdbMn(=olLlEOBnng zmk>7#<&DMIFh8H`kX11ag_r{YMa7(8S5RB(P^pN9s7BV}e2s}6v^)Ech~uU&YHqHwbHG|!P$ z35a&ypjMmj)O((wqqHG?{Lxf$gpm5f5H?kk*ert0TZ@CKr+*eoOmB&tE@6AiN z_3eSJQ-87$9_hfrk=#TJ#gOVO=!kEL8$uShAm+X|oKY0d#<&h{MTE)fDDfe3XnbJf z`n-k^#!;Tui{7>|>O%`IY$q+fqt}z$8n%D1EVk@PNFY(Q+2Zl&%(?WFytE?Fv$gMe zx_+569C6RFPyT$9#-QO7A}K9I*B8oA5NFAVp?^$W7zXZ4!iuA(Kyd5TX2<4H0560k zg6Ik+3+*hMt7)GJjuklRaJ2uZ(D=QJ?F8vuI{{tM_lnOb2QBw^s1dyna)flxzd!oq z=U`4`p_sw(!unzHvCG|puZK0cZ@Gib*{?o>gRoz6wSMf?l0izageJOnmiHg%v|aC@ z2!FcR2&s|$+&ejL`flw;vUqG)rbzW1#-ACrufOc_10|_?hkgB>XA9|eB)t3S5@0QFdBO_IwuQ7l07)OdJ zpH@V)+>FvB>R-NE7&&9B9ceX5m$YRd^?#t40uyga$&amo5mt8>B*K`mj7N-UiPLHY zx&K{U$$O*rx5a(=H1RVxR~YLE0YYdWHIdPZ^Utb0de;Mk?nFoc%rXv_e@1UOg&@1B z;yl?EqMy%)@p+ij#F~|mHelj4R7f&^8;#m5SS$})d zWe{1p@t}J6C9`VyFTw?M#Lh2c-uBlkQft9p`Jaahj1Tao45d=}FcIct8h#F7orMg# z_`Cqm78DrdfZfyIJeutuej6>6SI{mZUYio~e2h42;%G3zn>Xh9b~vRPS|(x)TY$K6 z4IP?I4A*4(%BEq5Q`2~XNoK;FMq5^*nH=g zRVyp6+9T)oPgH$jv`j<@`M7H}M^ye;z8hugueqg)HBqy+mAIff8{+l#1L^4)HMWl< ztR46`bAroprZPM6?gc0L3H1v34I!CW>kxZXaJ zZ9?P9xO?i}xF1eu&u4SYx_{=>w8ELbhz;szdhNH#BnZZ(UN|pwzD+MHbp)t!<@H%Q zTP8iYDc3O|`yQNqJBTgs+~V-#PK@xxCkO>`Y-@DEe{GpsWuCV5>`xmGdN>tq)i^}J zp}fT?lOIUd>KAw^_Q<`7ak(Q~{&4JP{>&Q3>$}8}GO69zNdjhp)qk$Js@JU>ql8Y) z>|7AWjK`D9HDcZ6DqErbTU={u4)*cF%wp-Vm!^aWL7SPnasXV^jNZC9x;$G_2V`K) zI<6P@;bAWuGmf}J2ej;}fVmKNO@Gj2>i-z$w7nAw`1Le2u!F92NgG_}#E@4uYRfx5 zt9Q!-cD7@;-aB}{hLqH!s_<73X zCh5aVV(J$T^Y%T*He+$+fQ#yS+uN-2(Gq;8c`8|O#qh>}89zDXgE}t-jcd-Fm>N?seyeB7lUjrqXn?q~kx$a3wXk89} z1Wz)@q+fxu9=b5-h)(CfsUI?+k|ewoql_%NykuxA#C}&)CQb+anap4i?Mo3EuFP zc`QO7@Akff4e1>Jw1Fcs@sJS$O=^UzxPd;*d4HLWljb~^hn)DK!X@L$5{cQUyY^-w+XXeGM5U$g3BaNuJG2D4WmfT8nVxft=H_Xl`zs`0ik-L1dcHSk^W zA@t9h;H|S4@p9PgFb}WI)L>Of__l`aZtOU&{PUrOrW#UX7?K8BUdUIC$EsNPOn<1u zr{L7Ae_Kn+s~5+~jt}gcq*N=cy2g@O%zXzN6B)Vta13u)H{OG5=)AI1sVIoadWu-?xsSO_i z@oo$_mY5Xil;uk&rp@{#h!Tt4e490GgE3NO2!qQ%t83Gwk4Rm&(HkSiiGRJx)D1Dz zQ&1ZMLLGcD1yMI)7|HY>i+HAX1D$?S7ObYQx-QlZ@wR?2g7dd_fnY zU6K3u`0L?bFdSB=!v5T;i7!I+v!1G0r1ECXa$u?m=TiLKWiBs`f;1%&VFe(wu1A4# zt>f>xDS**SG!^z%R<*V@V3zTB!fF$9on_HRRgQzT-;7?QixCwTefnSf@D2wl4%s7- zKRTOtsx*cAb{2Ruqkk77Gl`m(ep{C3S0vb_5+^yyC@xahHD7NCcXz&AnrDvOf_2>P zIgStx8V52y_U;54B(QiipGBh@AzGs^fRaeoNzBzCT4s-HKnH#psG zTB_N(GM0ua;?7R`bXaT%{I)!8+q7U39BWy`D&K%em6`OYaKc-vjNa~U_4)9W+FeY` z>>Tjqo{=Ld8jg=K@?rf(`gp}Fah#v4wh(>p=Q{uLUVnx%HY;Hsz#x{3izUrf({fnU zp>C_MI&G~0;+F$JHU#o1yd_DB=AdcmhjPDHbYn^6hkHqikmAn!((!o41}FVbsjc~- z{arSyg3U$SYfpM&34wF3X!~vVuMEMv1jeLt6%+w-Xr%j z`-W`}S${j1)@aN`C++P5x-r$lCr;+!htr3XNXaXS7{11ww|$xUbuR2Qpe|SYQ2YvY zDjwR2@}tH+?MztSqsKmdX1wma%O=MD$9u=dqX^`O05eIQWo5zD5X2#p>lX(fJDil| z4nm0cSAA|_4Qa&W-;sGn8+GJ7j-sQPGbcWC9)IXUO4KS?85z{EhEwWPL|y8`o|~P` zS1h2kW~O_Kh_#ekB}pu=`(%fO(d=;!&Msn-?UJDOR*0CK+uOw(R8i^Yw!LA=PdSh`_wF8s8Ejat@5w)u<}N(7CJg zge=6H*Rq|)x}X!ilak-R(4*X-1a&dxz&X=`{7m&khD>+*mD~Hsw(A08#*|o+j`2j& zRg=BsN2TkkP>3}86Q<80?Ms#`y*_!cZhvCbosQJ&e-drJ83HXSpEKx!YC!rP!eWBL z*xd8AOdhvM^Jl)!V-CKcVSfD-$HxH(_B0>0owX4|y@ZwJUR}59HsobNhY}JJgJQbo zK!;k(ah$J~3rA~>gs+Glqk)qL@M!YYDTbUzDQg}b-`yweeGxWq`u8KXw-)U@d4I?u z2QgA|*9LLhNfT}yaT;NNHvn8fqrXL`6LgCSAI_l-3ydspFfj3_9afiawk`hq-(7f9 zKG=3{o0@K!XcL4!gy*3M%E3mBbK0R&ARp;WGGRWujgwBM+C6`PptJ!OWLbH{A9JG& z$tKOdAtXw!M$N$|t6(O`1nxDo-HcD@z+`_^_$PXXGtSxOP$LoZO&J*w7KcSTGQ2Pr zTytI`+CmGC6{1#a-b?kIkmVE!*9j`JMlMDERbou(iB z6hBl~bW~Obuk7!VGqXlWoe=#bc~PvZXszg;=t__BLs=AQCK9_pjzdVcIai=iN9}*- zaUHb`VZll5*g!$SevF|6LCk%ZZTUV_G;`k9b3{{0@GEXIr@himj)%Q3w4Ml1E5Rp$ zSlY)Qpb&?y*!ayRZ@GrubMFBd`beCV2gIwrN%QxnD2dR>RV!mG-8LKhY{ERLS#ukU zUu>nBA;~V72;wE&b~!%>^82s5&hxfv#aBKoO~6FadJol-Xf(WRI5LehIcejCDiq z9J~Jheqt7IKox;~GTG@9=jV8OmQr;qg8K%YrZLs?8~8mxJ>&`;jC6m8CCEPR@9d%l zwznzcw+ng+m12MRz7m5sXY+&WZQ!+18@w=kkt83sEInN*`EY$%sT6$p;$^-NbKbFq zLa$1*lDbe=t3Uo^*IkGQJ#efNf-zFiKjtJFsLctq%1w-uu_$bX=l;}FW zU{z5d^D1wH#4qO)uT6i2l70k-;qA)39+;Lc&(ymt#*AM5u42y79EXcy@8=ylc#HEC zsMzIH195$mzEP22=0Fe>zSiw?-L7jpX6P5B$G0~lE>jkdB$kw}9C> z#be?`Q`nwug7Dr6Ivr(^M)@g)$?J$GkKcGgAGya^thfn40C37!`w7m>`{T&Z*%^W+ zKr=BR4WF$@=Pzz{pd+#B`9Wzc2%p4@fyD!P~Fx1eRU~>klZ| zn~LUpslnQ9F$8zO*nW@B_;@;kx%fD%UeBCB%PpR1AHWV@iweaSw(HT_>LyKE<+rTk zw_fyv9ioKY1r&W|t_yDf-uD^s8>05ZA+)_@W~>Cd(OG{sV?9;mIkfT2l;Q<|icaMQ zBs}kBCgKu9hP&c!tyCmU)11 zW81cE+qP}noY=NCv2EL#I4`zsC%c!u?LX*KSEK4H6DBA|fv~STIY;^Toy=wAPfyi* z2j3_niH+9}*V5o+V~Q@`X(*2^k4~_KdY_g;!i#@vd>W^2?J!xLU}VlsX^@`InVKN+ zVTSvAZzFdM@k#%yu;)xpuXQ8Zc9t1YTi=VC=F6XeK+gY6s71M2cwuE*q!HTsn%8Af z4oNp$n)mO2EjiKqNKz*G;~lKfwUEIBFtot2a5}yD z2%fvtZlRUe}YA5%TXlF!QX!(G_1U>smn>O~0ji}syAn12$RHD=C zjRrr0oq;3G?B^ONMKasZtkCR#*c-L@Car(^NFxa4an{gT%$yTGzMQBzU05T+M;P!s z<^^FjjdxnrY6X3D&$Q}@tW--Q7n<|et}6ZBm$0RJs^PI2Zb(lZzQ&ywdFn_J6a#O6 zy>#hjU=-1vkS?M^tBY(G%6%kwvgoMl;Z|8}&i)r_UeV1MplJ&14F1z&CJhyPJ)=r}5jgOOKPVHO6xEoZP z{p*PZns8NBYY#pB_jxF7QH{;f2OocOed{6JO#>d-TQcdWZejtPfiz5QP3(T-(xQ8O zJ%U`%U_yD1vSAxFR>yJm(tWxn;&DU{Bi5Eo23k6-E8dRgsA$(HK}DQhVp-N__#jr zR?_a-?#Vi5c!&`JjYyvO*rdGgfkslmmF%P~802+t3PPn^W=;bzSCWc$<>UL!j9qUj z5fre5OX0tesrBiKwM_Rz8Cw=^`-X{?h_Cu`(!`kRk+CEJpUVqbw+m;YTAf=OTRaMD zCf@8s&(`tK{wyNUd%&&?OACKs7zSbavvXBgdIi`orocyA_9ZhPY+neGRP) zvVFe?7p7;Ll<~s00Q)Wa z7LX^0xP%{w-q0qTxbJvLiA>xWJCN~~*#;xHInCAu-eX(*Ug$KpAlQGx>xXcjg6}7C zi3t{Nhh*KP0JGEVNPXo?*+QM1A@ZSk%FDbCsVmRf9(F7q+?>m3$HDa5P&IX@`+=-sxvO^N&T7ywH^zij2clE48fgm$4)G$ao z%S)N+T$W58E1S6F{$sM~rPUv3(VTO%ciH$>?YG*YEnmtOWwq%59|v$0%Ct4nJSyi^ z{8~Jexu>g$vU1GKvvS$tMxsZTNZB!c8kjQiV}w^y&P3%3t^a?^kJVzysGBBA#xGSp z7prHHr?V%JKTmKcv!h@r;jiz6WQxHKTNh;F>L0!FjS+yj5 zxs8^tuW08!>8Eiym$~&4ooVJDvIkL1W3YEazuzK3t?tsvq-FMxsYDbMdzqRhoC7bL4leqCi*kQk90tSHuoaGEoE{B4)U>b^ zAipIx?nb;7;jq2yFRs?Lp>fXKid@1f0#gdC<%4pTSc4qB#rId+*=}V5FU4u8(uD)I z*j@nGJ7`YtHbJ@AYUOC=>bHaL9RMjdumhD{-5xjcGvZE()h&cVi}32@P!4Z6ZPE<} zX5)JRwXlD*R!(z3_DAmGAWm^f4O;0-c@@!dvf6 zy`)<@Ec?TLhlhq|#HcA*+%8Lzk(wZqRqMXoaO8yLN(fY%yiz5;z#oPS^@_LlO8iA+ zJ37)DcFl0!{@d{dBjw{Bv`V2ywqh(#VD~kyx5a<^lf|Yf=4AGBX(1OCZuPGVT|;gp zy692(aaIQw%La4z5H~J739%u#q&7@MVEhS!JETMy11I^kdl;mM4IzNE^bu~bp>a5o zJK#2OY-+D?_*lfLM_J}yQ66m`1I>7#`hIgh*kX8|wr5n<#)^i@zZia>nuH>=4thSp zetUmB-}H70mV7Xd4Md6sxqOo3BtBdU_E0D&@Hx54E*t%td071+Q_AKsh}~K?--q}k z15~d2m%heKo?;T+phZ+LsW#a*SqbOmUF^*|DPY$7raCqbi|{V6H6J z+}*^YmwbM7@nNxOlt>@%%Pv2~5Av(j<*0whz5|Cp^yv$(hD25d<~-)*AT;}2Qh!`^FgtzhTeI+)q74fM0ddy+`urx;qdbNg{{iAX$X&nQt=ni31C|MaXLNY z7^|N32d&c;iJ3eLF=_5A;2YOX;u|WaByt-Xfm+bcl1q(LBwbzrdGXJ<6f)e6+0TC< z$}xF6w^vFPe@cEivTZ;TN^r`}N3wysf#)eB2u8EKbA+yJD%p?ldG7W_TxrWOMh$n0 zesO@T!1Y8Ut@@q`>ZCNb+Cqg=n)>zXV|}?Z95?%dx$0IvoS${mS4B*eV6yGF?Fv8Z z`szCUjx2l`9%#R)(N9UD+4nidZb*NRiR|7lMBY*q3dwA#oLBO@9CNQFL#Q!3;a(g5 zv1q6=#gGFYs0V&tRZs<~$K%<(d*-}O%m9ZlaGsCiYfC5cU_euHo&ZWkx~o|l zQBVy@3EK5BPRx~z8zsseEJm4koTbtF&#_9j6QtBTmw&WD$V~J)NmE7d)qQ`YlFG>? z3uO2x$M|$xQaCIH1`_oG_&pWbn9VJ#fVlJ~l<%2+yS@o;FZ&dn_f15=DG+~h35m); zx3(pk^J^LXd6^yiT?&hY&G>@YtZ>TZDyAJhm(Dj%FVFNN#!#F#7^ie!Q`j8fmG`jQ zu;t!CD%DLXE>Yo0Ur$3@)!Tmo(07+ud30XbrIER zuisjS`4>&Bwdpl%u4@Mi%%MS2%x%f|_$xWAwsq zzmqefgXXQvK4cn3+U$UcQLOa7lcve{NR(p@`T$)+%>-7@0s(DQwFoE3mbyf=13`Q9 z8Df7I<1ezhy#ic5o3)QKrWChv-+iW?)L*M%J;<#5C#gSTac z^Nj%`AwYk=>`~r(m=k{k2cboduHT6T%FRRjZpf&t<(pYH*BQ0eLTB@Q6qLr^p&QAt z)ZD@B>;xxe(QFNcWa=vK+zg&ZmC}|3ov(Hh~d~Wv7jX zE3CW=+i!1rlxb=}La;_yK1jQOdXpR40LW2-Mls4`*n^W5ml6jUlk%fO3BUMKoH31b zEHtAf40aofprg{J&&(ARxX=!xk}`BFW}3Ca`2nUUZ#~Qr5!jdc@n;NA0)5tY`t2l4 zo(B5?i0-jo2wi_4AVab6V?$&liyeB^OceV+c!#3P}1666TCJW=lbGEI4C8$O@N|`fk?Hj+NUom+X zH64~DD);mq#s$-Qh)ZMU<4~%Mq;MfmfYEzq3TNsZ4MczKraGkJ3-!UGqHOMB($LFB z8{eWhWnq~XegS7NW#GDvfoS=Jy|6|rG;PN#YSm1uUYXZ(QbRwAt3dOyAT>_FGPotR zvD;lW0`I9raj+@mm6n@D!?CVLVwxcZj2KR$<0OCHYF0=M}$ z7vBI7+&MuBV8PIkpImGO{7ta0O_tXj#y;*hs$xYZS%K+s~kUkSgOQDuwy-bhnIpecvctxI#mjShID_b)aebDUs|UooTmk?y;lhpN5-!Y zzy*_@vnt{`CJ)x0H@bbAD?GdBt@tA24ln0RpCQ}#p!!8a`@In^!>N1Cro0p@=%Bk% z3khRAYHs3OKpfOr$u=_n$6d4)a^6@jO^VSJI3%<1j7(v+lH`XwZmP>W?d}Fjcogk7}Yb*EWYB6nL=3G#FDhps|)iM5uP_5Nl`F^ zl_L-Py^$p3cGWl<^X>}vsJ#T6@J<5KsBM3wC%I*7W0ue24-_yaj#j7WE<$O=eII9; zfYV2d0gHl_38teLV&3vuK=^aNS{5WY3{>gVOfuU7 zTU6NE0efR6cB6@o=k(}9n$dqB zOJA3iin$1?2x%`=4BZYM`ym`2lf8@xe_*Bz+@m|Ms-F29baY2GwC=6V_4UFE<*G^g z;=6S66oT}jTyvO5tef!yDbZO>{UMNano!lM?Y{~&(iQ=)so z_V*xs*kb=Nq7G*R;Q(lyi1W4MJH~%2`}Rcaoq7)(=nr#^o-?c2L<>PF(kcp?h4xTC zVq02z^zSW=Anh7wRP(I40)zM(ad8~@{ugQ8=RHf0QLCcGHoc?;V@XE~>i!h}=FRf>9xSuYdbBmjnrFWuZLj>Dp+#1S+ zec`8dlk}|{zW$X>dWrn$ZF%IOoOBxVUwN3Z@IVIX>q?av|AYv#i;cz$`twKhrUInMuf6RI9w78~9$bnE^iXO5!wwF!~0N_DW!U}X( zJzNMt2m5t{BEVubAY?Iv9TkO1@5R-S=W{%B2g5<%Gun99UDZ)M!*p+;zXbwe^AU|DST&{ROx)MGr^o($_N#XS2H}mMRlYm?fxFrIoEz~>k zSzF_(@=D#Iz`Y)?xNOgFTQ)rR7kWQI!G<3`8^o-rF*Vn$2(9T5j}R|6Oe%hj4E}gg zOK#eiL3H@bC+!7^Jehwh?W0k9KFxjUw2Imf@vEt!IEn6l0xd)%Q zh5~%Qtei-#c;4o}NsOiV*ko2g0>YJqkZ^pVnklDVH*c!ayt{uBb?J*(`sZ3r^=SPQ zQlDv}rsp_eeajAwH3R3c6pC}Z-dg2{heX%t^pRZZ{7wz@f1ffbXYL+z_9o@(^MSAJ z@4LfA9LG9l4zne;rY!0p7afC%9%GjH51hUjE6ZJS5O*00)X2h<#8Q+^m@WgJ zbZFeZ$hCj*2?$PO15Ke2wvg=Q2e`u!PX(B;83HW{h)TaRCIh*cwyy;3=A1&%PbGQ8 z(&e_@PzpaJPz>BTkXe@Kmlt%IV}Z5Ts)cr#U+k=?4^#@Oo#9S3s+x`m^pLvH@|gZq z&bt^{gptq;SaYT%ugO>kikFUd*Hzl#5OmU<1W|vbqK50hs=d=Mc!)Ebk3< zwO4U!g$%<60iRFPFNm#V|K*|3r{pslrjNVIV(blw!;n8P72qEXuP#g=3VZb~0ao7A zy2oSsij|wIbi#>yOsI{6Bamm&qSK|GOxs(*TZL%hT}+|Y$thyt-+|pkfwEtqGTzv? z9<+ZQeG$Z@C>;(;k^Q|cdI|m#QF9Z9zxEI&&syYpIov|udWKy7IYoMUnk_ZkL|wY6x_HMO_rVeiIPD>Yr8&Zv){B z2`W<2-%Ygx?rTSFM+5>I_M;^O_UV7GE`Ocw}mu?8evoUO^%>EiHSg^ z*0d)nlQxQKD@U3=tOwvMTK=O-~w}rZX=JgQF zr8=X`wkH)$<$35tW@sPP@}+<7l$o@%F?>(L@=mozRfJ1Apo=Z|q_KYkG;2v@=-VkF zgEp$z;tOGos#%+@FF$ar2dC(j*`D;N%bKMh%(KoRfApjYwpr~QE$|n0^fg@coN?GU zs-O&R8rD@*p1}C3J8N$M$psd>-S{*$aiX)$#UVdfbzW=7YyHb?CW(Iq=@4KhVX5y_ zA?VnmQRxU5za<^;m7dwTkGzV*W#qKnRrOBd9fijk&vHXD_pF2CTCO<7Qq{BzsLMmo zBUe{-*>0p9l}z0AlFyId4(swm-?{pih~E)Jql8CM+5^=?QL`2ubQbrFR}y)&Y+9{g zvKg#v*7nM`Pi|Pp$#8$@pQakYzTI|+x8{S!oJcHv(jeU{KBZe!2E~X3m1eF@&ib5s z%fOXlJ@`oI-!^H(PiAiOAi&2@aa)kb3}KKim&70*v(1eg%z9ynjwy<&ARH7_afNcuwJn#T)r@mH0qT;RG_qF4{TM}$lbby zxNc|1U^*F5fOCKIZscdPbP?AbqRg|FqU{VUjhBQ1O=XAphkWPk6w_$7Ge5nW*R32c zD&yPJLZuf*h))fp`Bh?ud4TfGA|^!EigJ_;a7++%2`xBBUSTo&Ja->g&bB}Q=AHq8a$&S>C5X`VKXD+R|h}!wv7^dY{F&d6{%S1{FA_K2=;1 z5U+>$p4>g7rfayoyEhvjTM>6A0~Rsq@D0x^4c3(U5_FP4dg&udmsJL2MiE_oY@~n zhecXV@!5V?B2M*l=Di#P)-Wk&;!lD8QyST1tNnk+dGoW5)J*M0;(a+NZY0$+&qk9J zb^>e4B=r^O_2R-Mk1$&z4pSKyXT*t(v}_B8mRIOLf&S0mb!c{LxIC*NTncm%Vt8cF z>O_ezX#@IhD@gilRYBgiAik{;+oMYm57)171mjAcw;K{z@Ol4f&# zniYR^yfqpvjw!Kt)cm*4ls@>wM3(}4DE5DyXq?~AAA;Io3Co_lUa&9>BlCpjPSv$Vi>6M zy6vAXQ(W>m81BcA&J4T1jFkbn2V_zmAjE&SeUnhP{s+3Fg=j|tNC6PYec(I)hb)$# z{`fQr`m8{5&?vkj958Zh%L$`OBuxK|{gs%A&4@`4-o+$=$q*q#|A-Oo9xf~juH!RX zuZOKGFN9WyoB-KAG6CuJsIo)#nQ#GwK10=&!d{w!Y-XIYpG3yP&L$=PUsELlFo=JH z0r>EMn_W#NNceHHvEceNbdy9*=^}$P-5yu^?3VRruE%J#-}$AbP0BK+m@J~tgv3GJ zhO>`X`357V`CKY#*j9WImYp|Kv5?PLoCaHj9G4OU6W@&HqXJi@J5%JJjYhnRfG>~b zX-jPQ2z~G?0D!^g${fiGA(RBK8*TpVl@O*4rD9J^OEoetES z%w{==>2QLAkQ=D&C6F|!<({LN>Ul&P=FVU^pL#zp{SNN0ji$GuT^X|z^*(>U1vdx{ z3<;p(6tPIsk`=N-K$>bUc*}2tDE)oX<996@Oy`zF@g7q0vmf4msiaKu$aHYHgO?z1 zeuN?&YrflDeN~N7*n4UnJ3WpPDnVlD&3u*UXVNajL*!1iQNr{4g1?getf}CeP6w;* zp$qxXwv11#*j|>-^^OuD!{ItA3ePE*|V@;*$CY)VRXaY8~AP@>jS6A1nsaX1IF{_Z}=3(|dTn3|IbC z3{Y#%-P$)vcOXo9lMA&ObLw@y1NB~1$##Fe4_pH8fo=)GIc6&%YXpDV6vB^V`TmXR zB338k^2Aud_+Z_TN@V;GP|WejD>Yp$stAnc0YdtOB8o)@$|bNhUl#+yLs)}(v&JvP zAU+GEuiETS0?IyTE$Ey*1^lQ)C@wj{DroWt4}Qh!q@gAn{4wzC={n;f zJybD!JqIG1=6(!5{;+S?*_hB=RtQM?0u`(Hw2%mVIv#PSfO%#3*!>|=>9oPI2zBMo*!#Q{uEvZy%_F<(&1wnO8= zq7xq7abe(Z^LB)&VHcR!e8MQBiADa+Z(xzfyWipx+vB+1=CT<4Dhs@?Dg;kEaliA@Qw%2Nr;1xiz zW^c@AXdb8@_;mRj(OFfD|KptlFGD=8m|rHE$}=;|9>o!^Oy%?IM(>H&)O?ldaZ)fJ z{;tl4uoQnN$vf#>M~2G<=a13Bpm7%?&|R%KmYs_9OUsT9{9zIS`;_*x?>0jeIFIG0YpgawuzraQ1WA4epc zNW*_8wC+6TKk#6$AVF_PQ_oLXlb&$dVF;|tZZVg24HYPapZW)=aV={>jJJqH(!@DM zTvP)O?HK|vnv@9BTj=;?AiQ@Y5zM7|m}-7K1XN!5{T*`1bwk22_RT5Ksr~FsVawf1 z`^i>xzghFND7z8!!)P(UNvNc}1M-rV(?ow5bTd=8|ELf_0LM1ej|i9&MdnPLbjRHb zQYB`km8bl@xWtxdtO*{B#!O8HK_Is6(ChdzLN6Nq(r+q00ji6%o{LhKDV=IznNGDP zgi+~|BreJ^!G^nPZ-`5qqxaMZe6Pz&7P&5I12Tr5J&f~Wt>D9vF1Iu>K%>k!5G#K$ zP%Q=OMZq;}oj5}WeqFi({96b07i)Xl>TD)6q2-{KZ^~zp8mz;};te)-0aEMalZny0 zoPO#%)blH)W^iGR2_(ec;1(o0xqR6M!hU3C%7?z^;6%_{60p#n`NWJY%Hk;W;+EAK z@}))+q3PqGa1<$2fBi*5;uWQmWt)FlfsIT+X~K!pMsYfUW;L&8z48U$5F9Km5>jhb zN3uj)F$b*R>KmJ9Dv@Z-o;EcnE9p#GE%4j)m_L!=7X-Fy-bw&g(1e*|Afx}Q zwdV5q8x^$$%7~lksuO2LDxED9+h;iQE+ux3JMkOU6MM9?zrLt7l58F?gfu(~U!x0r zBew=i?Pz%xp@y|YFI%yA1vr)zJh)LF6s_x+A>oMm{BBoQY~K?Dri1|6DOUu#ck(%@ znkD`%hucQ^72=yVXwx|nbXtBarANsNkE z8?ky?zfWGVHw5#cQI&lc!A-ALE*fvV`ZX{&f=O>q)L$M|Uz(?pV1f08uU}hbRUlJ= z&<}8V*P7FPtReDrBb9|f!`6d%)>81RZ^g@$7k`o4OkrW{eO>q^+>w8ZE{80HCL!M^ z*a;=g?R(mJURS1~_dvz$>zbBQyP_F zOW}nc0*mA8cIQ&*VZo7016uG|P8N^(5_f+eG(%BP2+)ypeZYQ>h@vr=;#~^HB!8oX z6&|>-st!+o0I>6NiBEs7z!J#DgKk{KuvLAKf2BG}W7v4T85oS#99w|VQV5VFjtqVn zU#!k#1%+gHRpxZJ9Oxf2PtO7hD7^ixw=aU2sao+q`4fLQCnSMf*J232$>vr0 zpI|RZ+LH|cRuZjHnU~U9mVJFM)M;lQTviAdOgaU;F>HbX`wUVIjTl-nNhk@gU^1Pt zWsE`M_Bo5-K9*MqgBCIuCnDA* zv~_fvB5KmiY+zT6!Eh+ZU_750r2jRo-rp=ewR?|lIc|Fw>DWIuzHy)Ne`aWucuk4* z(6(TWbG=iMmu3U#P?IA?DO1gnUe5mT%DXSt8MG^W-F52_J5Hzs z2?AAox7~m6$ws;RZ3()=sCR?o@Czk4S>%)aP2F*gjI0qY6DdJ6DIGAv(|m4kbY2Vg zFFNnX=Cf(IVodpz%S})0N+Aw4i(9Krcc$R=w+A8A~@HEy5?WJnhcD1**B zKoQvY*H%6?HJ$s@BP5e+4T3C?(&X#lV{3}{=WTx{%10{iZaUBM>b5B_^i3BbU>wRS z+GF9EIOm5A1@fM@S^$B}-Ol{E2P%hDu(zQ~_kDu*+*{87;AhpXQly-gik9qHiPs}x zacR=~D0rUJfb#7<9`Q=va-_T-bXsl{qMdk+@!{0QY55N%fL6X$s$DU^7*I$TJU&-` z@LPWxL2`4X&h4)^ACiVr?PIeA?RK0=ymHAusU~3igps^6Iz4nnZU)fWPn_7m_hfxA z5QT>IlUYD<5xY7p_Mq2z>4XH0Cv0z=+YD6+oKSsrgMAjk>ryCcY&J^9aIbSvO|5Ip ztXX0=ta8TBMj<=o(_{H;MfX@4!_h6XY^#5g$M^gVn_}@n4bmZcTVzY?98fGsDX^nX zHb@HKlmY5=J3Ze#-|36ooORV`=N*L4&-t z?p6OQF}APwKl_H#d29d);nRV99o{ou$Y8f5`1h|74RP%{rzqc`>#yC_&4J*O(d~aB za@kYx6B>#F?1z%y+;M%KcB1nxo`^bn%XKQkCr0-PQIOSKP%X@`IUJXg;t zlJ(`RUz1b(KZ^EgcV4=y0FYLJ3Grsq>j$#9?_oYz%?=bC@0~)oguVD*rX+{7_WGCk zC}yV7VcRBYduxquPgK6Uef435R#|_mB8enOu>4`#+ z@clAI2J%TRH^<2Iwv?=F4Pg%ZqAn0*Dlf>Nhi`9{?A4(qVPwkuf`oCgqV^pw%;Rr3 zEMrgiH_phn_F4%nQ$XS=&%FNu=mp24mmyps69G7v(DyDBw+m@40t*8%I5L;f_bwB+ zBXlh&2?H@WG?&r$E))_tH8Bb=Ol59obZ9alGcz_YIhPUoEEEMZH8D9bm*F)5Cx5hh zWmH_tx-IU(5+rC7oZv1YxVyX4INfM)*Wm61cM0xJaMwTx?(V_ib@smJWS?{I-!~Y@ z{B+i=uWBs@P?9OCFbJ7~je+7|kP`z7BQq~RR!$jgYXo9pW>5y2JKGpJ09YBBnb}b& zDMcKBMoyMskf@OpkQcz^WC2hxaesp3LL!)%c~B?;5Q^4y9OHr!|zJJvxOM}(ncoMU>8Sg zOMnr`6d=tg#|V%IyFiSV02(j|U<|Y{vM~dI%>b%EO@Nw;n6e5$LRmpgQGbP&5mH#i z+0G8^@IPEcR8-X@=mDZa@~UD0pgKK3LQO^W_g7UQ2*Tf-9w4s@@&E1v3H+TdC#EW- zs--B#!t{F$02aU}po633@3#NNjS9jH@P`^C)yx5G`&R=1jfInw9WN7;i;D}RxwE4a zBiO;5(az>CeySFhjsO?1gMT#u^4$Sw1N=)EXOJmGPA3cC-vxf}3P9G<1PF2j{!S7H z|LwGeNC`=T*q#1O3?hWn@0vD$hXWje!2f7tVdVH%uB@V>EWp;t666E~8G%e7flfwF z&W-@Xzig0Cpegm=1c3k%X9tJhJ>>r7a`<ywr48^md&l2B`&6-${-6FJP?B%7@`h{(aF{6 zujt?PghXX|0o=@-02Us005fDr#XzPaU|U-VYe$sd@`+kPbaDbaxH0|j!L|m0T|gfH zhr`SgWNP+XdsAmSCVw@MrM)vyO7wriAtsbRHglj8fEfU^2LN47ESP>b{A*r*n^}IF zAwqb0*n#Z;W=1xSKrc%(Amkg0hojLaAi&AN8R+HlPse{E6c#Rksilb%WYi(Y1?8{m zQXn%hfah;Bgvq~L|1$(Me;rj?$T>9ygKXRYra&_kCV8+EWPc=R{{Nh`|LP^~Y-1yD zWDBJEUz+}RppmVmjoW{Q{#S?w@Hbl;d9Z`6kGmQE%Xe@pwfUCPM_ zGXFv#a~mLJTmCYs{XSAQklBUYKbF6*F8~7zJLiA-AY*A_4FWnk0@(jf144xOFUAn{ zfAa+}DQQV-tA9zo|KDTtSDYBg1Z-*vG6%47Z~}}R9E{vhm?5*k%E1BfV1W#_DbV$= zVF55Pg1}CY6o8$xlNZ1Y?11unLpeDBOhUg+eID(B>o4n0GK5HAXWgA8J4rLHcR_8}UFoX#YWwNTYuv7H$Y_ zBj6t|B-8j0f-p7tH{yW!Ou#mf5&4&eo&C4P*7hIum?6zf{{bP|{;3Yxa;Cq=_D=u@ zBpxzNMt_bL|F*F~Rt*^e8zb9))Pd+?_6M>;h?!absmt~oe)>nF-{D|qhkq1+gqZ&W zLYV*Q$PUrc!p+VC2>M3^#Af*q2%&EE4+v4w`X3M?jmCkL=KP{Yy`@<#cO5IG|!2TNBSX2>Ir1!9MM z{Oc#Z|7bw@=V|cu&gAno_C@js8EwK5Xa_ur8%Ku;>uJ>aMQ?J&YKWjOlTxaol(s{U2eg*w149r z?Y?wYu2vEIbb3o{3kg7NW%X{UzYwZ9Uf1!aWb$ zrKgsl>XLu5*wIqYUa~a>Z>p_?yK;K4d8BKD}j>`LjZ_epus))ByG-Qp@T7sXMs#gY2 zR&hpg-^OO~ekm9pM_&*B^?wGRbV|&E@)F$ugR{tSB>Uh-TBvi>Sv{vLjuk%{-&l;{ zXL>HxwZZM9g)wbS^lWv4VUn0~Jr~;5qUwp!p~xoM?wSKi4EcK5)#`Bd#jwYjxT_yh z&siCd39mgl?t>hKW1=SHpxy!zOckiGUgXWI;wbR!skYX%zW0;}Wq%VcI6qg=1Pe+M zuc#N=KlCaQxNV;b;C!SqSP+zuF>7KT){C5)M}!wVKcT@*RPiwM=H&^EQB;WHf4uV_fP)&7;z4_Bq@{nQJ z9ABKJ@I*{SU-ukmB1z`ovoUBgguh%)PJLY4iiJ7an5W1tIJtp>*OJx8K$N@IhgB+ialyXL5jw|`&^(ji>7*;g}m2y9D>O+62NqOb2r;aTRf-4>eigSMSNQ|F}$BADA; zL!CDQW#{9@F!0tNC3`)j>5m?58a~ccXDi_Pg9)jEPW{*7!nuvRIQ%MobRR%8>r;ca z8D_`OO=|3PTT2+e^jNf3sO`P%oOxP_^=QHeGcmM&vVY;@Z&!#xU)Ll)6>e}F?^F5& z9&EH;hSD6GHR1`gXa6vcH`Y}++mIK6LVO02omM=2OYxcy;xydYd%p#cEQ>-=ulS)G z)A#tZa0>H!J*@Nr?3s(kJNEjmsh?;t7vX18FzkXHic;iTGT5nX zoA=OUi+{0r#G@^$iXY3x;}(dvkfJ}*-91D~PAa5G(K=YBtDARCPa|Bp>TR9YE%hhF zW-16x;wKCdN5TGztk-ah`C5;M7m@j_H69KN;zRJ@l*e$AGV{T$$R+Bwp$ypAk$j?n zeu7QSD_j;bU||4qysp?e-rqKC!PQ_*y!z?!eSat#Uh*wlY)5IjUZQEU?>#KOo?Q04 zD4gqSh4Qwo<|NX-uUK}XGZ(644J*RDBp2@)Py)XBo7{#xvDLw~7as*+89{yj1g8)? z0zK^5|DiE>TOoTkB22*Kn0a)IM5i&dCgcNU02lS`(D%Dv@iP2To)T{-d#eMtoIWBO zn12j}M4LJE5t}nu1@^J0ak&{&HNkx1R2o2oKn(jAx)<742ZhN|kZg;pv>G(9mgjJQ`fN?JaY@ z;peHiF|Q>coA4{;TNEH%ub%&m23jX;>BAY zJ?j{qRg#x8UQyF^V69G9`b@Md`gsU03qvaPQ3>h0MC(p`jI={83aO`Ee=v0chkvPE z`>8#E1d9FVvu2&P{p|kJ>d0pg@%ipc0Xb15vEZc75)qt9@7sVp#yW1)DxC-+71u`b zL-R8l*mR``ecCaCy4jBuL%9R(c^b}pOzxC=sb|~z$|)Ir%5GE921BTdFU+AZE_Uj= zTY||?`n;h0V`(`q9R}g0sm53C`hT!>=a(gdXC9oqJVzQa3Ckj~42#EZRBGT_^DrH2 zaq}URi=&B3C+HF7`no2iIvQhvsU0#kowi(?6suuNd-;vv@^Iu0mB6aryd3?XHX$6k z2AbtflbDZDI^6h9d5Yq(uz!C13XrT0 zzFVlmXdP(o=V*;YdIs{t_rINFX3?;%)7MgF36P=!%t!styOYa@*NDaV_;shWC^-Z)PR3Gq0Uw~U&bTO2w-01{a*->(nghkZ= zqh-yHM|3Q^dVjO*N`FI|UQ*4@Z|q^=wY+;YCEDxYExzmf!CH0=2 z_)PbhYB`;xDqRhBPo;xneu`0gN*=WZ1yAbMR=GcKlS6c8dQ=|>R}-e;{UYYd<(c;x zU=jjiQh7L}+asuhKNV|uC_j7OwC8h|+E{L>N63cX7p~sc@PC9TJOdnxSP(HD6cg`W z-74?htqu+$^CHe#v%tIwhdryd64x$NmDHzxeC_uBRx2dHPTrSlnr)8_$kv>>^BJ3> zz)0rmlz94yW%=}8#6V>w>GM`Xx^LLpVwWVaKszlsD`wAwL32Ty;5v3~6BVmMH4>Y$5Fr zov0ZhUf>EfV61Ieep2(sk@bI)Vrh9!i{J zpV19HPA$06CVt$RXy`dZr&#T{AlU527Z`43k_TQ>6Ka3zJ#PdH7WdvPdiCu~Camx9 zdiffnCMZ#_!=JcIYPCX{oE`{=MNO97Bgw+O2HAeOIlxew^MesHOwr-8b%VcQrsur# zF;HY;BI2I#8h#d_o*0R4*0M`UWQ8Wde$F}W>zpd2o#QMwIHDnMn+@sQ+x5UhR%*&` zxvvqfF;#!tYJ9}^pN|rZq}+Kz<;bN>Lv^afI*-5_HG#1Wm3VI5>~9nDYQ6g&s41XM z56_ittP9lLm0c}Xhh?-Do2+H^I_f7|RF8AQR&;@SX;CbP{VpK)a1N!+)y~42{81oq zzJZ1UsWCHIa%4@p+|yTZ_?dUO5JUca8Lo9U_i2A}`7HlS`}5=)=ksKPp6M34;h3c0 z2jO|@rJ!T%jIV{YEw9h|)uW%Lx~CL}S+WvuVM8UmyI>Z2+PnhKLj0}Vx3auo`M!_O z*mQTLwLejY55_mLj;*!#3YSp7+;TBUF%;SBKut;Mq>hwWwZuvsozl*y&miHJDe+G6 zsndV!uZxxO#AxEo_wF^I^bG_6SBj5_xnfOH9wPSD))~S+R~)`3CB-gK&T1Un%*kF# znbZq_eja3On_aV>q$FYS=DpQFThpc)g45a%#8Ht$T?pWbv9AJ&(>fBwXvfUKZk*L& zechMw8a&0A`)c)q8G4m|?V%FsbLvr!`zFA{X7$UhC3gxno|s4XUa zar53Ap}Gz(4|cSm3wVrD-RveW#Jhjz2(k+EdK(+r5DT3(FvwFP4KT58iOG8?v#%0{ zhPEdnq@$RT729n-3Sr(shZ%O*IQ@Q9!WhLR|Fh)im*!Sea^!l9r;eAPGw0y#CsUlR zg!ow}ypzM}QN+%ajH8buubk&OD41SlbTJVg8yH>>C42d!+l?3>RmOsxQv!b@E}1Oh z00y;0$qK#op7*;J4h21VWp=MEhx)0%<&h%$v&p-i1k!d$+C7Y2hgZBEnkbp0nS!%p zl=QUv2G78slsC?K)(CZI4!0FPvzEOfw3kx3@u~iWQ<^63-Oel}@f8U?e#sY{1oOxT zWO0FY{fi%WtsmZY09-gLbsm4eqcVSFFvPr-dX~p4%!1ji>(8iog}QKh&b9UFSsOXZ z(3|p@_NtQ|nK`i>P=4k588vc`^R7y)&ak+piGG6`>$}~}2!my0O7Smup5%)nJa5v1 zykhgVfMNmMEMp5DMV~9J7#Qq=M<2PGHsTZ*KGh(DMr?5sOZhohbRK^ef9i-~V@#XM z(dr>e1*1sah+XLRi+(IhsCbDrPNBgI5)KbxI7}@Xt58*q3`gCkM8PX`9{lk*x4NJI1JN7G1PLg`0zR2kcuADZX3W|+^g1v`b*sU zFCNWb+MLOL=v3ZT&NNAGunK&0V@dOhxz`%PFXxS?wVWB-xT1f5pz(!^%ViaPOY?Q7 zi%x~O_T#3-*{zrFvOS!8L}?7qXAb7+)nuDTvNKvr>)=Kq<=jGACis~7+fVN}EV=hM z-NRP1#P4|Co3bI#N}QWY+uPRp)yu!?z{*2uhEuJ5GjgHZaXssB_X=J8hu{Qx^X?`UQx)g>H6c~!aBr*@Ysu)(OM5C=%DB;`S235o zqDPtfGYi}SS&GfyRJ3Thw-Hr1A~C^rAyj$9o>+$$yWE&CJW7&lmFJs!6xYiSrp=4u zyrD_}5ruy>ye}~T1gdu{2=d*+(wvGRe*T=sTugf+!hCSBU+e@mSDM62t)gX&4_0nA zj5lcomy{<>RLJ)>`F{ix!kdn&985>-`F{p@XfjL%A=&A73Ll(US~R)B+JJh-lGVh9 z<)`(llJXF6me1lRnd;t-9TdmOEs{4XP*R!p`*?rjHzO?Ka8PJY%=vkh<9KriH68xs zXcSid<-ZH<50`JYI!2hHwA_=vjSr6mHp4^Xm>Nk9&`eNl+en1>-u}YRbm#n?z+V7k z`hc8mc8yF=8G5VQxPBi^?K*+3wU$?4_Ibdb_N%vEVOLf`=N<_3D;Z1a{)LEfb+yMO znmB*TlIjEObf91BLd?`eF{7%mf3GEb`-$;>W;jZFSum_kZ9K4%ZQ{~lk{B5%6DHp#>*UDU&_*i(fVLcYDn}qk@4fq4hjMqipeU}mXJWC~|dP#ptdy578&<0G?n;Pk(uD8na9&QT1MD*pC zIW9IWnM_0GEb%t{h#T7=8F3kPiLSxgE36qD!H!sqmr^l?b}4GpQ!M_E*%5dZd8|o~7Is9!dnzk5U`c-y0$q+@ zlON#C06olJOg@62)KAz*#h6UCw$gu_)b`sgzI9!AbV?A$hyq}efJ0AV)ib{ zgPpS*3NHzkB!fnhYy`VL;QU7d|Qey85r1YXfY2NACBlK6unjR1!j@8X*lp$_p&q2*+ z(Lt(^eZ`Qele%1kuJmht_%)+V-BZHl0&;F4v2(wX5kEAJJ^-d=uTc=hnP4&Be zKmP8=vEF1ydCKOWw8}%((A3~xe8wGIoxi^Bb~1&o@wY#|(GMYFcuW4p4^Gj}x&KKW z(Gx0$zOeMHo3Y-~ipPJ=Q@sb2)cR@D(3iw83SP>3*C;*0@}_23P~3Bdx4mc&_QpwH zmMF!aYU=yR55(X}(}uK6G>@JA^(a*pYdr37R zBtFev*CMkYTFe$kCd_=*EbT2lCkY&8cX*AP{Waj5F6ZLF#jdfY$d$sG{Iq~j+Zsaz zePKRf&|}$a`sQQgx>#gN*OYVL=~FdwEeB`LI7~8r0kH=cWtmqi#Z{dq)}_)5r21}p zuMNNV%-~5nUaNl~?$g%Bjk7yret9>niU|lCN-ehk#0j6#q265c>`&GE-cIM~F3m0} zwIEeJ3Ys2SPPA}Qg^f!T)wZ=Y^v4}@vz>AI=YkEsYWJT(2{vAZ;~ZhUY!*&dFpi2P z4|fu+p5o^Z8ln5Vr3%>n!jsp{`@*>kzb-Wk_WT; z<$De8LF6LiXmVb~3f2X5Xo(&m2;*uXQvHfrW)VNLL{dr}#kC+krkgBkZK4&OX1>Qe z?(-7#Y~rVcu9_a&1Al*mfrKVMM3=~gHLC;mug^PBW_kP~BKbk)XVS;&uk+^$Xm3Cm zud6pa$C~xegHT2kA#zgALx1aNM)I_wv0d$E1(P+O&*=R(GwQHYq zkxcVrI+Hy!6fIv>g+3w#P`FebbctHd(7SpcsV09kj+>)Blh1d57GTO})Bn{xP-x7N z89BN>-8_5~RK4pR;tcr!CTv(ykx z?)ZQ0M*C7@8(9@V;Dwr6v#(KMmT;=c9A~4`qvBRj`TWVY~)o-pt?8`sCJ8{LJ>U1x>1b% z$QoFDnzqofwxW!F+SWeBzCQ6AeV@6co9KVI;)b4iY&7(`$xvCh0qo4V>aNTm}#=AdYQp)pTdQ0-VoyPTw`x?Bu zC!#luqX&@bRLIE~+pF@*j2DclY7BoRll%;b&h+L$B`(NdqJ0zyjtA{xKM|BevF4Yw z$EyeIX;XDyy+5fC@(HOM$iyzh>ZyW1uz=%j2nPnSkf#Qzls!>I*TMR1Vi%nJ6yzF| zu~j@mV^E~!tiRx5nT`&~Z(_lnP989=VjQ#RD}u$>2o^s9*9t1De9J-%^n!mU?nm#c zNyBiT`zW(U(i=qF?g%F=GeIUt9P{lw4LP-0kfJoXX}oGgHeq)NdChkh^$Dj7!R4nH zr=0E<3}NJg0ZAC`IeI7y{Q%!~o)YT@LL3PTIj+KFX;;u^k!6;2!ikkFa?*mktQmj2 zY?Mi0Q#m4$QN|tzUhX2%Xls9yvp`(!+E~Ffki1F3yR&Q&ceQoK|3@51Wqfx?-Wo@_ zD=h8X;O)1^wl#$hz=PP(L?Ll6F#`RPVZ^~U?QhkA>jJzj()f}dm?2S_zQ`|(qzL#v zt$9qLZS~T{)eY)BjEp+QLv`*k@3{!5l~yScsBwL!V>z>HEzL68LWF;4{f|?qUNwnu zeWoL)<0Y{O4@MA^yt-x$`cMaK4L6e~p*a{*H^$Jpan8uayh@3c#~Tcbe zCYuvaol}j(MkaV^MTLKxv?fR#$=xH06}AB-vWdf6<+<(Tt1jk|owUUZ1*g2-x>IHw zO$1UKd~jVe7F138G!xYnu2Z+)2XYx3&td{1tF}6Q?G(veN&P`N>{D3QfYf);aw2&u z-``ftP`HPD4oiCVD*_D(UK`Lt-e5!l;NTpAdn^^<%diy>=8{JZqQpPrEXj|Sj;4>v-Fl7^$RH3xb9 z7ja5m&U|SIXB*Sk~Vw6CIa`f|oAJST)3=8zVp?riQtG#Tbf| zs57c5fJ1@@oYnvtrT^;f@G>O?SP7`_^)5Jf?}!K*Dt4LTKQT&@G2YYLTR}?QZngT zji1O69}Q!vmfW#17Bq04NajaLLGz}q+&`aZyeHO)^;Cb1v_R|+YP)lbpj1(HKV0`9 zmE3p_Ex@J+u2AZ-$NN4%BQZ*ZyrOAJU&U(7pje!xFL#G@zXbqPtB57u3eeEK2*;LEmMPQnNCudx(Yy(Y{;cc+7qE|UyAM_3wj z++B4I34ir$R zLrL;gt*($_?AhnO@WxuVG5xjGCNEH@e8;eRrN(y~-$%{$pv&DY!l8W(%jR{QOQ6&I zstkV~ZjAqv_ycPozD<9Nl@O-(Rfpl)vo4A7=9EebG2@R!8r9i@*Y9!w7`;xED$;x_ zhthw;VGL_yCNIcGCJDhil}_$cVX<;bV7ZoNj3Izzc8<9M8WLvobVq$tji$y{)7`wK z{v~TWU+I@^rTl8)pSZ0n2#58d7NN6+eFLT9onl{Q_7ENX>wjiilHPokrrT|gGX zikFbGTAZ03^XTXt1f&V`2p8P#X^C$&PF8>CNFb1ZOmU=>_0{<4Mm`Nox_FArl!VtbO0YHkr<8i)8{rut^Y z9q<^#mu0H{jof?A&LkZUUDdUE6d1u9!R3e0|h2ReAZ=e*FHY$8!Lk$4!MP*)syV)}~E=k-Joe4Z{=wn?D$Cb{PE|HU)}|5ZG~Q z^(d*SE<$a7vU#k+665oyLPwy+XRk{>($XncA||i*&)=|0S4V@jq$&0%$b5f1((yI; zY}ymUeDB}9BTXB_nCOk$kDyJlelrPk_|g?jgRd7ibZn7Fqf=`P4yC^iRz4DmDAZfV zHrCTgn6OHjw7TTi_{z*l{>t)=#3*I`xXtZdd@9}y-3RoZVQ{Kd+-kWFb$9-V*$iEOs zNIn)tMXl?>T{?hE0%@)ERCbYvFM zu=Y)Lu+N19cCRgJAvX!Xdm}7x4@kGm87r1%aW)(CUuCcou&oC}wi+LT(r~yq`@z(Bz&A zHz9n52=enYVd)$!MFBeWNY#!AR2D%NbGxhs3 zA_Fl!_=Q*;7cY`(m+`?70>PR?gj#*!O0}Bj;R}B<)ule6BgxJNEV$$PSym#knXW&( z72&!Aii2L5|qEfFZfE16Ht3%}C$y0Uk8i&S~q%oT!CCxS# z`Aye5X*+amv5E?mc-$1L)-=9LGhl!$|5vK*q5>!SQ;rB&j|rR%{ZGh8yaPxt7Pwj= z4(osPghU8Ce7Km{4-CWOa*AVJCPI=}DvK!MmIs87Ma>m2KGLJM9p+e6&xp~=r8T=K zr9}%~QRSAKgNXj3V4?MvdQS1u#)!jSzw^TdBcU#i&vNY6)J8_OU!$!Q;d`R zNu-7>N3gR&QHo)APhWoxAg; z<-yP?@&g>UnB5FziV=2LVIuJb4orLe#9KI2Kl{?#>mpZ0DV2z0_8rR{gyRKPWflCe#B-sE8jsMqGapud*>+tSHxw z(wfxp-)Vb)fuVThs;j~GBoA+dzo>o63Jp~B!qcit3GY?COf5RVs3=O+#E0W^=?~hO zHt5Ba!H&zh<-lvBvCHH(9sbzZBiay4t>6)4XsENltBK)OLRlvTjl1eWY2 zRG#FqMRydrkiM1ylevF)8*yQ>c1Xdr&bb%-q|wjAAbwaYxd#^MQ?qCOI57I648({O zEo%Op=k&3b^~4%wmZ~_?-C3PEkpsR7 z`!NW+9NVLVQdL|}IczQ%mc%%T;rCGr7WIDydazg~Txt;8F8F;A z4Lp&#znr3ei&XE#2;9xUS;!Y~DJSH7>@;@RburP+ErR+6t$c{IldN27j~|_w9&%pa z6_Q_u_z#QLR%dd)(4xvrH6c2scN)61-?s}z07}_wH``95C0OC^-#;%l8Nl&`^|MX z{-ZIySV1TBpCbvh$$9MKsD^v6DFrUJvh5}V<%3pdj+!G#KhSs3k8&^~_R)UsIJ_!w zrUfGXge`VD>WJoMZ`cU z@6`y(%}{6=OVqeD`JMu4Kl+x81ms}D?@jbJ^-{lSjW0cP#9d^Zlb8gedc!`ms& zC9I}OGXk;)Jeyk`7 z!e2I&e>TUtMoK<&R?ITpPfvYVMt0>Un?FyK!YzQLBDVZUH)^LZS8i>mXsx9nvSPBCANONe;YHCl#Mg{*UZvyIQ3b4c& zQ5$*Dtv7`pxV(9HC~UxEbP)c3JU7dv(T?wYzLhp`nk=Gf9>ABPJN=Skjcai?H4e?Z z#vFg!I${3sUioWp3YUBMM+T2RyO+_yc~`mcu`S{{UccTvg+u#z-NTsTGu%WLxK4_9 z)lH4viF;q4196UQ(yOZ+QAb7q8w>~)1eNdflqkQ(Wn8ug2F`v{e4_9Bn5h@fAITo- zc(x?{=Cl&-NV$(Z;5@8fAotoDg5=yl9H@FCbB5(uzlm~NHSw2sTw9pSSEHd5^$dvm409y~*rZvLxvYN! zoS)DT##SyAH4(HON>|!{cBR{5M~2KgWRdTw5*$P2@)f^f<^z1+{E$&Q`qHr-WJ?fE9I@uhO z^pQfT8xsU2j7Pg7S99SNXUb-zjPQT{uKT8oG|OVjDVXQSBFtUq&WRzz$XOcS$dq(A zir3=k_icpkba13z>^)#2X=}kva&30JXDJ%pE{Gq`Pc{QJQjQ(x9ldNTvNj>;+>Ai3 z$iv~|NoCZX(dJkqE7EJ`qT8rF^J`uIuvpdQygm<*mUW1#ZRt4vO|}2SK0TdZU;Gq94L1u!o%YRrOFWP3vO^pMK(lW~ zYT@4}nvndMj^g4WC*LT>>ePR|9(5ww+Gk8N?aN-HDCt!OjDhRJ1JP{XoVY#)R<|4V zQAy~seIqdnVs4p`z*RBsUi;PjVZ0q)Y?;Gp=VCh(ka}xukAFH1Go#d4PTId{kkm#p z7a>fC^(Df(&VChM6QiarkU#NwSusNVYbQ#fkN&J_5S*u0Q{j~kXYqgSRR$Mfo)QEnB5T4N@5t7d;fkKUHh`%Nm!d0wUZXoWIl_le>LQZ z%#kL9dW_W^g5+_MlQ>%}$nls5F8sL4d0)ls`t1gNMNc$QGthsE!k69PnB|=DVre^0 zv7*L&$9%X>i(MPfUD(b#zAPW6)I&RijJY2sXHvmw;E%CU^ynp{SCJVrv345|O<|FO zw3}*-d6&7Tzv1^T!jW^$2b(_5L!yfz-(tO5P*r>}sCic#Gu(|U^@CX4c8LN{(Y|^;B(cBFWn)P5d?OJ<`Oq)=nOVV7#)aK_;{FlCUW8Ag zUPY@ya5m)mm_ubbwDkLuQY5q;Gtp#}TZ}7_(>9IetAxUg)x5fCdJg5?S0TwaDKVVc zhVK$sOx@(5M4iRL0rjocIRE`FZocxWYmAaZoz1)gJ{YPv%guIOomA;P1Gy=-jSi z=)EYcGUEDD0h)Z;Vn~tgO05*dh1%)lvJ~)qZMdBA!}S>@S^BAhK)Y@4K0Z%y{+=yV zH$8tHY^fa1eUe7cR2O|F`>Bo`T~lsnwtxk{=WEbN(d6UvUg)UL)9J{Zql3nqMJ6QJ z1&`Re&t&$s83~vthga{jJ7zqIkHhHH1uVLI&fbEVzZPh?75Xy1$75QpqnE^-e)hO< zt-)DJGYZuXU5LVJ$2+DzEDB!9n;PR>+QxrEC+6(Alk>oN%w9(_ZmposiYK>dZ7U(1 zSuH?(xWTiJ<-!*M=iyucD|AT#$k z9%xYuUM*9L|7f%_gZl0+2On&R&y|0mhIkRZ4|QS<8_M~a3(*uh{6s#gRLchbneM)R|BQbkmcf6p1pu8p z8JC+|yezm9q$5F-A)0QrFn2zufeBFD2OlL%J+%M8ik`3*Nj`AKE|x8;*zrD5S?H|% zx$DMcow5IGbd?KhRoA-aF2Psn;JdAPPQy_{E9)YHJ31qUro1wJ&@^>lW(DaGj*)2l`qb#L0~^KA@iOH z1}{fN5-M4RtH8pl`<#EJ`?&Ee&STbu_!LD4MzE>(oKNxD$J$0p4Vo%odbTfj$M80Z zwerYai~Gbv#OM4kWAWr%tFzPH10>#!Dg+Oz)}BsYczKO3cloLTjGxkzV{<2|PJGvG zf$2r5=W!gvVRSQm@Yy>>MnrDCUJt>(S0{!60{0)HN7koOA}fE(il)k0%3zE0Rc_#Z z3M%ll&igFt_kFRy4EX-4xq*3#DZ{<4O|{C{-m`tce^daaj}E~BYGzgio#Ugj-Q3d! zO!0==_JwcmL*>*}vqt_U8Uihf9&{sZ)QfS>G!BP&`Kb5{K3j;;iWI;24!I}X7x{Pd zK0_J-wV@pC=R|)7*KrF6DptljkE~i$SIP4|LZd3UE1-3I}lAZzo=z6c5cTSDT~%dPF?jshw;)pR~BP zMcq*J=G1raC0ZC~THugM&Gj>t6FJ^aQ&t>BMfd%nIHU|HbBq3hsXK6^Gm+`{ z@`iu2xYs-^ftg<*^>n$QaF;>fsVRhD6P3uTNqXs^sa)Q&LkjBGksY1FMIqrYAWu{B zyQLsreEo;Li>Opyk(`sa67e5@T03-CoN_HdbB1iDG!w_DGQMdEuKtDA9yy9!%H5{g zoC6fg^~ShIvk-5&q`H#h^6<2u5IKJs$C!otPN(>6#%Pv;LaAZ(Q}hb+>(_z0 z8PTxX3r{52E?&CYh*+o40Rg_q>0W_u?#R2G%^+AsirldQ_t1wpoBpA5vb}`wUEePP z$VKRLeD-xiex-cpT=ZIsn{ss0wy2Dwf>nzYi#Fq2;P3h+PRpP|P)J+)9Yvy(#Y=zo z?V+5v1isJMO|Fb6j*wS^s33$`zvuPeWz6++7g;?DPSf)8Ve1-4`vmoKpn^qH+e1|OVR zQKgyKIgy-=Wg-#T{~rJ}K+C^3L|%V}ys5%@f)E<`mRCQFnk&D{-iE*@fX6MlhVJPB zg|GnK?5PP7+IDNPQc$~DGnm`Ov~@`xG~!xm}EQMU67B>SMLeLODn%Eb;wzdF&p?lGz26urU4aDV?6Bf{Va~S@PqBqVs)6yiP|| zd1n<5qLzhxFCe)jh}&&IGT48S(A&vKu!XL?j#~1Vi%3o}NOXlLk(_TPy;Lr5V2Zm% zxMiw5dodiz<=ed3W(11_pp(Qr20oDsjR|Q6Ah>JLOG8W&PXZ5%#gt*EWCz8iAF_h- z7#_e10tlY{Y^-I%a(p^cfg{v!es5in9l`}d?>YM#g*M&rKU}YCd@+AsE7Or`cMg9o zj(({fyXnm7go6NqAcqp6qL}_}UxZy{NFDhXncYf=1l?XxxaiZl2wn!RAp@n{J7gkd zow>bATUyVlFEI&+qqyY%m>|7?!twFK|CJwcF80cWW^caJr` zN%ieN6~lcU`Ar0pENQ9qI3#1h1OHJFr+wq{ltjr^*2-Ex0Q7X5iEdWyDwZML2Q zIcMtn6@|2`s;|@tUJJ(K@!;fUYQL)+%09{Uv2pU4hPFzzbILj{^H25!b|o zHR<1D#}wGixK$vh3ned;A&qT=Kp=@6hP8g35oz;k|FyrP|vs{i1v2sU3JJ|LiTknX99~%U@j-dcma(90LXcW)V69 z#2~L56gbDa;ElbWn9v?>iKliHF05HAMVR!QP>w6(9xHPd=pceFF$sOEb0%Ks)Tbqn3GTSr4z}@5?Xr;If2|HT+5BcnW%pP5pbo~LOdLYkN~8_qxQo` zh!_2)7uW&i14%^uPcAKYVm7tT=toe606y1Uzq}(0>L0<jKkT3vyp;st~>L+ypTJ!=p?SQdpR!P!T|i%O8*+k2m0`v zWz%6mp-hWY_MN2{<|ruZs`D*vS|MmWJgs~xnCLe}hudD!%$7@mJrQjY_qfb`WE;hWC;aL-cCB3F@ZLQ9I( z-~~0@O2nz_?L2AW$)2x>RUlr!-d3%%*X1Y|g6S9Ygph&J7s}ZoiBAWQLTA5}AmAj! z{(Lugc8-1Gw#U|W1*@^xYgaCHAx2GdUn`*enZPQs39Z-0#^E?E~_&k64ifIQ&{GRfIC@kC*x||RxZBMW#bLh{H@#&ph z+bS!8f?v{8S)mVvt(q8cY{L(3#u1$}<;6qWWQw&!=pidRFPF81Kq4nkT5ebmY-wr| z4kLi^UuHof%}!~lA^`r_da*~&xR*6#e&Ja|PCE8jk4OKPCWCoZGQn+mb$)+|^h1wS z+!2+`;Z~t6$Wk^u(;gH@;(oi9&p5@NakT5{+j#>LCvk|xFpS`YvI)H`uvW= z&{FkVg2{(D{pNnOq_ZsR!sy_==W5gRpX+fI~CuQtqyHq)+-M$ZN=mo?Lpk5 zg&FaI^uCSkbK?}8AItH*bMedl>F)dOeX7luEuTo%8T6caGvjfxZ$4i>HDa@+DzN>D z)v-CcQNcgpz2(76aLVhMZV7y`sb>)mOQ8Gi$=$;{<&+BY7Dhv5DaL;fwnTrBN`UA3 zT4n&_j-pDC)i{=Ui8%#)Vkb<1p_PcbDMY_eXwvq%G8)+w>P~{Q(E7EQLlc}#zrYr# z=##t(r5|?i76@XK@jxZv56aOm;6hZjhZWzz68v_tm>n+HRAC^0JgESpQe5Pgm z^ChI+>o0kfZGcMp&#{JSZL*qm zLma)XGh=-jRD6w+Y=3YA@q9hXT7;t*>PB@{Gfa?&9OrN!Xr&W$jXB0Ad3xyhd~|DV zJkITtx!bVHVoatXWqg(FcHSwIWz)SfAGss{Zp&9H76IAU448gA`@8IOym#$pd_yir zfMnH6+0aRly%~RLG9;w|7vHc$Pey^v-ZzrBn=uEK^`G^Nxqi zBZ0k^hIq?3_e~Ofv_Bv7eS=S~kr+=sLZkVRgaY3As?G%p!Mkc|Oi!vqWrstiKQey1 zU`mF#=!Gfz)|2~nD9V^PFQyC@t4;3s@Y{SZBT6t8J5qGL|Y4t4bP9DwS>O=0^l~oh;F5S;Lu2L!xUPMg~ z_u*m#w^)Di^9v*X9*xq%di}1QF~DTw;nWQBY>)4HNSU;~-3l}SnPOxqm|y$--wm9q zhzZOtc8Mon#MP3rB?Ugj6a zieY6QnxA}LEXAzt=6GA*V!8egIkXf2o^GE`JQWQR_JIwynE;vnniI%?%H&Yu;o+or zqIq6tj*Lbj(ZZZ+wTcB~%`~~*c48KZrF{k71VB~INs{M4C>n@@NT3(&dmW_F`9R^e zCrN)n{VO^=&eZqr(pap_%-@KNezx zg(%Sgc5;o5q zbLJK0cGL9Cst~fN62Kv!k)(wAA#08Yw-JAN6klfEJNj>@J(ZQnCiBCgu#{|cG^@X` zgq)`cBChtq=u|=aNJQb@XkY7V{f6xa+p_4F$4>cV$a}i3G#-*fZaL)EjL3hoa{9h{lL~WXRjfcgNc`CD3z+?8Kd}Sa97a2@G~X~0 ziG@|6D8~=}3QI32#TL@JGqk5%wcMkhs1}f%L>Dhh>=ht@4u}4Uh5l5A{CS-oj|2@y z3R3lY&XY%btXU+FiFK$gBOp^~v-N*e&SHegpejo-viJ^l-h=uF5-^UQ zoLK|*MWQ>+F&YweT-U^Jxgl4fwb%NvzWJsK`L><0iAy394-`NyVg|{si_m|U>Lly} zu|)@ME(wcbb0BEYeVV}MmtS{Nu=F~gD!kFYIb+u})=L3`rq9m?Nd^4e1>_>MHMG~+ z9669pJ;rR z^D)QypG_fZbt;OWf+DMZr6Mfnz}~t_%{NeN)0Su#g~_Izs^nIHeq~e$fPKc6c!&+! zcl~}VMSMX=b%RxcDT*m*(Z64R6D=_KPgay+DMU9+xMT|d%8?FYS7fs$L}~{cWIw! zdty4pFlN~+z@T`>1LwkrQIBzU>lt{-N;SZ&ddKR6&I9xl?uE*dd&RoD1JUcdRx zssXsH36||*arx}=5526}q@3ZsNyipaComCDEA&S9)g_|nm;Qf${J)HuXzKZHb3km- z*+jT{0GUrzJN>)`jI6>&0=@MD;Sf1{uS#}N4Dy}mgPW?740}v>8Pylhc|67Lg5&9Q zzopgZ*g`%Q_~RVNUoU_CYDsYE-!`zV*%+sfK^eC2G`xIoO_wH|qs zmlz~HvG&irrNt3{c9n!b+0Tt@YjEWDuPVmU;kH5MR<3QcC}vg%7yh{a4&TIugV$!fJOm^dY=eIA@GV(tR zp@V7)zY-&)azP!`f|H>)JO58DYep-_zHXjS=Yx(JD8)~Iqepaf$~=LJtG`U3C&i{O zrMRM*EXP*xxSFIoLi$KZ-|Ro1@XM5!<|~=d(Nc(W0guf``I8)|(*0d2&I32&>RLFz zCGq*fKcsvqQPApri~`{MrqY2r|NCNA#-}ND_T{S~BOf$clXN*^8Lm3#Qh=VvS_MtuY4@0U3C-V++}J`YBd1e;Ykx44-e2-t zuT*>>!jOVMozyOZCNy|=%3Xeo%vr-3B!(S~NT4J`GQKFbeN7O|)|#MF;;uFbIlI}px14W(OXL+ON1oN z)E((&B9uzz-(r*#b66$)?S~37*#VJepfE%lLKkIeR(I8*_{rkvrW9;^H9m?b=hi9~ z;^I87a3a9}zs&L-`pX+8Q7fqFL+`*Y*u5I?n#Aa!y=B3ktU+00%p1UCTfX$@63Si< zcz`v3hdTICZ{oxWvT>ePY~S}-rfWPS0&)3wmJF1c3tLRh@4Ann(?*5>8s7-avqRb+ z#$u#`meK{uE3lbdB=-3bB~eL@5-VS-Y@tsURIJbJkNgsJi1q~Oadlt2f0N#h1{OTwNaB~yY1+vo z%w(GX4b)2GL^0~UnJf@^_}G&leezspi_w-4OwpczCktwtiG34%L3ku{*Ie<9(#9@- zUZasCo_zn(Ik^lAnmr8TOtQ3%v!1O@YKb`+O|bh7;hfxZdWdrkYUtLD9a~cjh)8IE zw}yvx?IOvkH@VxqK=!+9``OKKAtoIxffT5VW->SEDRBG_z%s7VqWvwDffVkprOTMe z4&l4idJzZks%MuO5bs2j*8toqszUy^*pf(5%Ap*yA_!khBrbF_j$O7RYCo_rHoz?? z0jFir*#=egLFI5@K0AZe8mP-sAo_8@CUl<@QMA4WZofe9_4cOBtCW-uH%l{o!JWtP0( z{$SB(fFMryfdb9c>=fQ-+}1p5)dd4Tp|iWXn~b2#B8J{HoD(C2*AghMdl;U7hDwyX zS`NKu=%4E!yOcIkdl&=TzC+iVs^ryYs2uvU52>dd7cL!XWKalRC=irt%$^(7Y`anP zO%(G^}XP+AOTQ7pK)T@zaHR;4s`D{5JJ5jR2hC7 zd8S4VO-*)eF@W5AN@1Jb8Rg@DdC3XvqeTr=V8_@wplitvdXhpL5RXo`zFW3IMmFUB zc7P~V`?aAD$CN$uM@mBt_%qud2S7gmU1!;Gm5)7uso;bh@57;A!^^Se6c}QX)J@E4 zY57bVF+^@5AooFXPK_9RQq`XykUK@d^eTEieqQ9V_IzJ%aCvDiDhXPDn+#&t(ED?1 z)LcM@0Qx54abIiK-X_J_-#uC%EyDej7mo>#nk57tH)sj9UZZC+k&<(|m3&NqQU0vr zC81uIXk_==&JH6Z*DM>N_fE`L+)_a|U-)n2aN2(p=kCWXCZ)-ef1ukghoXEFfdN9A z%i*OpTEj4i-w}K22VT#Ae7iwVYe)Ezt9{a%a6_-g3>@!Q4kxz1OnD1c5;G1snQ;?;>{mNCT=!ZEI)7$O z3#V6$6rtVH?b<>$odF-u2M9IE%_WC@HFK50G9?ltwq?Z^vTrA{6avpbg(w>0jS zf*SrwgtKo$Qq~zh;s^-?nOInps#VvK(db7sdgj!a@IykTlvh><`L60%a>z+ZFbB7KGoxMHB&DPnrW<3 zriiA}r0lhaEO}^KP|Sktr&#_nr!DT`9YYYURr$*BFn75iD^#$T<%j%1<-kir0(iI) zJ>i7rgVF=84WU7yzNF^k=Xltakcc~7_W>BU1hQu}C~pv7?2W%8)c8h|0r^lH*_OE7 z68atEVTOT!%BY&q!l+(D#^t2>p8+$}6pu)&@&q~go>r5AzPEF$7q{CVG_wX?ee_80 ze1y!_%g2Sg-7SW+FZmzNeg;EZvXlZcg#VyBnQBXO32+tQ#P%Y@ zwFnD#x}sxWHZ|H^Y1+&0u}RBkK@2c58iFk18t6PL=Q)U$oq0+s!>*A)OR2A+$~Nrf zxEa{DLFC0w@h-|d4+`WgycxIyTUzr@M%pO#dtZ@ZZ`p{*{cwT!K<^@0R03~+ zenyI3#Y}FD&Y^F>eOUIbHM2##VZo%DJWgal%bg#2Cc zvS>Kiwed9uz#pHWhxuw&M=c=ZYjZz!&z(~r5SVJanmhAmWrIvTuO{WErF1OsFom#E zDlvi*eh~8x{s7=y2&z!(zJRr4+ST<5Y0wZaO3I0Ug~QRA8XA;z%mBqb>ckNYm|s4B z!Z!nwc^zfq8H7cPRwR#-P-rYcpnHRrd_QiRLD=nET4@@3jcX`uGIGFQxo+$oV)D;d zZ*(?Fj;um^t{$31lX!Z;b1aBZ$MxTmZbxX9BGuvNgF30DE3!F!iO+KW66!{;AF7#! zF_(yRGm@T*_r$TblF&(7w6l82n@}HrAuI&l&lG}D=_5DM0=@KY3Q_CMesKZa&R*u6 zAM3CUfN0szC45&KY;jxog&b#sI>GWy+MwYm0s4gJ@s1zPVNYU?>=^&`RTLiL){~zc zq368@E%IMSh$TJts+o@0RYBJ~^ex9Vv9lz18B|~BjTmEdNQW8k$I0aWEz+`oN_MMA zrStpgt8$?`$WW--+OLc(g3Z(OtY#5!Ab@Nsz$wc92(1g)rO*(x zH%w@q%B&Xz2pIIPjrc6O3ue2w#B}f0E}68~Lp`If?&n+ykAjWfRT{;Ax))vu8Gk1Z zdm*8mIi4C{Z55J=pOfZx91vn{tNr-~Q15c4lM2HWhYG~sbo?H8AjFljKC6ebIX_1? zGbRryplHbHqoD8*D?|17XZyJc8idx;=XOZ??@9<*57bf40UP1MT021!Zs+8i_$-ay2X@dIN~kq_p% zGeTl+_#M?AWmz7KtP!Y9{yzz0I}h3hLm^pN4Y~RRz(Jtmo2*v@s4zLomY7<0wi)uj z-9T%Qd6fofb}#gmvV) z3D7r1lM!Wfv>5#(n30}r&)QMq7Z`>N?*a!mA!$24a9rea5hHbLJrcsInU1wi4~wK_ z$pTw=d75k+1G53Du*PSg%LR`{MdGk$wPu=L54yI=?|psABe0^u9DvUWi*N9~_C-B2 znO{y+y-`+w72XY~^a+yfo6aY2Bo$ucrhG#c#WwCjM8VDkz<){4=O`CW$Vn*K$fH-w zUgm&_LI&r!!>y8*v~9wx&H6r(fTVt?9Ts^g6v>_=F|0FqPpnLNXNWM$Ywc*Tg@pnU z(IG<(5$JlR!61XOny#sq`jtJoj+pzg8L@@mRW-j&!id zYz;i$DZEa1?Ukrdt3VWl{DW#1RcG*EW^}jP0ArbBKk3=Q!@YfV>whtH(XJ6S+2Md- z5grkLaZm@eF`@6X6{=eD%Mn#l>9~_M$MK!XS@&I|tST$`tv_h3RkvO-ua9091WHs$ zw=DN>>oW(9$w%eR3FZ$7Ifp|S$T*%Uk5xu-%v%xCxKA$XgqQ>DTuwH`9^u3*TjG(& zk{Ks7l&DY`qAWlDV=#iW{8XQ+#DQXpF$0Q!168_*cX({NXw`0?gVX%^*ygQ0to9im zC$!(~+JcKA{Cq}1;U>9N(kDR(DTulP(r%U#>aEAJtI)+;2qmmFT-a-tOG(o{Bq+vd zltML)Q$^M^?+AvX3udkDn&S9g?&}OP?usu!;|u8Q8k&Toqb~(+oUiF-6+Vl+oTY6lkL0j^RYm;&r!Wmu$cs2 zYGokfeae)QuC@WuJkEi71$!k7VFy@SJ?{B~rPM*&akY}e4)babwu52&PYnV8<76U( z#_7tQ=@`*Pp*W)~Gs1z654=y_X9ZJ#DJR7o?m%-Is|_?Gpjm-CbYCr$T^}824y^PC zS*y!bSRZm-z4-OUiEWH%mDZ|IK{zuyZ%ExA6WPnt6emnm2` z_k%b(T=7(8a(4^5Jg}ZYtBleSHyGpxZp`v!kf(`EruiDe&q@>^X*o5P$B^=WMFLcc zqdrBvC)Z^t&bnK$<$MC#%GZ~A3Hq?ahYXHyVKRlM6QBijVp}lf(OeVDQr%7Knb%qL zzl>YO+@ZtbH-#@PN=-$&s8>|&-=4r9DC~P2vmKbsM2Yr-%Hy6t=TtmXlFia>y9+wp zC${X?Ev*Sq5)cA44GKN-goX!yV|}UF{x})2nSyibUij_g2s*}6%gpSDqdXZQqijj4~{>VggX~5KK%iLF)YN7rvj+voqQ?7bvag@W?QUvzosFyn9Dg!v)rJbZ_7+T|tI0f{(?Uk!lCS0vbNs z5?H>Q^7vv%&Nd0OG+EM_YIv5*B~oB}g;hJq$w2`)w`^HNPW0{=jRw6~n`XF{Usyp( z6Ihh|^Q2%`ib{mxB=O?O%kpIaxAR_FmX`w0Ac z4rRXVR7x*@CT!lr`*fhY>)e20g!9mMCTcmePfI_54@_ng$vF+#S2vv1U~Ur6G=4?d zSDZhQCu$%Waka%3XNu^m$>2Z=;HlM&Ylz|i){NDL=RDef=893_!E z#7${`v#&4W&9f}2F`uklg9C^M!Oxg!tE^+22XB=o;L39#ezO6MTnXxk$_R6=X%3Xs zJ1|qW$jMw7tvg|%U1NOU*kOL#lNo8(k$DvT74N$xiKF5oISL5r3t+2RL~|2!@Gh0A zc(1!%$drv#DTe0e<#V+G{o4%SE{!0mC6WVwpw&P28|Y$C6~(a{zD{M8uMBJU@NI3k z)a1*5i!iK`Wzlw7@Hsm)=Pc%c1T3b5u_K^U7N?}S_8~-N7auzD5#n+t1yxE$CmaV< zgh4|?5tdJgaBI{*rz&R?zY+C*Ku$&-9QKUx#qJg=Nkq`jtY&=-%0l=;jwjnlXvP|U z!tb;s_dOi9QI8;X05zLs9Y|erGi@6oZ&&6?fD{qAHWRK8=Ev+YoQ=F8n%D+D)W~^; z8<(m7hxiSThLb^f9o{@Cl*`Cw|6@>ri` zE;uu7HQ(X`$B-n6)bzks42-*y^NwGCJCS8?&Kh{-OdBt$H)UygO0G5&$T0ad&PDZm zf@G4fMM?MPx8|Ia{lNq~ysC0Cmf!$JKY~Z)Uyhus4rhatVY9$28Y%QB9adO2>g9iE z1d#rur{aXL5=OlWhN{o(d@Yf_PfEpzUsAF7>_=YST3Emt2*ovKWCedDC=I=T(}SWA zw&W7+4sXbkSHZ&W7zi#1#^uXYQ*+mUM}ux55h4|M}j4%rgTp!R90)Rx5y9 z_R;W3kdBSsq@n0rTn3ukvAN2Bano~d3lh~T(tI;qWM6jBV43z>8W?}temSS_c!&~L zFN#-dyz0>RhS${vi)GoEcM1(6vC4CseM@7mpXkPI{k_$T?z;6>Ug-)=+xf-y;1nLo zYumT?63&yX;*R3K=df6dMs?y5*q$3;u)W}9#76)!JW)-F+QdT84CO?BjlIrBLR$bE zS!H~XL^Siwdr3Rgb~(}QR84s1{s{m9TtGJrBH=cGR$yG!blVU}H~gwu1gv- z0Fu-Fz5bHVHh2jZ!G4A+p4@PZhP;_>BO{e_4kVv94(p~Uz?RTX z)Z@|k-9zNqJ`SPoq8Gb=^FS%lwe$_Syt%g~tU;9y;`9Qf8kR12h9&E9+f|U<1GO7% zx%`k%o~lGFUF;xwTiMYa8lf+S+~`^H^oqbY@@?(0#EQBnYjf2qjAwB+OvVI6YIm1f zrh=j4e203y(bvtez|`lBq+t67A!RVChQXjTZIVC$1?#(cIty`sS9{b5!CWCy+)~*M z65Lx6f~J~dqL);i0X!jivA2ypz~|*f(we+tkxxQZJg!+?8hPAD zEWi;yC(-B%(hb3Q>ac0Aax;IaNnOgw&V=Dgn%K@!A~d9}Chv9}Ml-5#NA*2l24U%YR`R zWz_wV-Cwmt(WlTO#^%XMGVJr7DjGb#64S_%59I%+=~T(liF-;v3=n(=bzHpl*{q)W z+j=GNM6*4@&>&TCtb-4M(^=g5&@nIruUj7o2jXKS6Ch@PRG*d0ce5>OPh*`6f$j#= z+NUV4T!NZp@rm2)<8uheo1S+2lM2W8!)!8R@h6Gb@dl_TzAZmI=y(C18XiEhZ`Xr% z7`)-?A(1EQuxtltt0T-fKWT2V<)Dj6Wafi`f$S|PhBeg+I)o%p`7kicSPAc#>GX@B zbuInMOa)kfahBGu0;*c3I3iQaE2itr`@(BAw&XO*RVuO}9&NhU*OMIP#Ao@-W$-?F zM@l|T;K~$Ax!^1zcSZnpF`dxHA`SLSVNLL-=G?c-Zm$#hj{M`D^qK#683P>QuSQN&LQSgSzKNA2g7;ci?t^>JA*d= z=~j#mZu*4SwQjyS)83nJA7Tab(Pr{QUIGiD>cAiR|05=}4kzF=+;!*Ym$g0XRF`J~ z%OAF}MrWFF06FGnXf?vup2?rpX%NELSYUsDP{CdXyFsaUbL)0YB@Q(k3Q9&%I*j$QP1vh)A|QrvuUs^UGg{B!-LV@QMGy zPY4juS(LjOgptV2GLJ5G9Vnq7a7D{A-}ZIGK8!(dR88N-$ca&AML|!vS7;PY?H}5U z2#?d-5Mv-<{`dnn$%zx_{7)pq{_047#HqD+l(OJdo3tSS_5f zCSiK~n$h}lc$PL=zfMC_V&bP8_Tz@I`@Dx=j6-%Vf%~mWX#`AP{%3dA<8-tS!mZ>8 zh{fE1i)hqgpiuzax10E2DEd%-XN=d^a@w5A$ zDE5D3s#8iT^%BAIyj$*{A%GPbx_Il@0hH5HYcEvii5gLr6NZ)U)t<1E*AG&=D>RH% z7iaWX-sU#5Pr`heS5E-Wr@kCa`?*UN=U(vgcDh7G_NALl;f`NWr=WVG0mtVWygWPkIr-lVRzLRKUN25v#54Y3$WkwbQMdcQ`N$3B1%>c=~!}9 zhWv?>Z}+ZwG}iOClA$GgpY`h)yMU%bUbpmCeY#aV!!8gzo+dGWOE6|h{T0`>nvlL| z8SNse=sSFc8mvFHo=ktu%}eP=kzD6p0t1m3pV^h=0uQpaYe8GG7YOP{Z! zJf3PvQH?@X0WRunA|VfIz>qvf()YiVQ6;md2-aEx{UT0F77n z%NhXrN+zgJnc7y>UhP(O-6~M1yO!FWwK!>9L?FqHfXhJ6^f^}h+fY9Qm?YWN7!LUk z3R`$kTjzU!5zkc!z-)hmdEae+f)|D)>R!t#t%CS^LIaTwnTg^_t=< z$En_w$k*Ml+yVz7ze9I9eu-Cuq{@W{ZiS{Yyf|uqASwPNId+|A-o41JjD{YN++Kc1f{s#o15RUMy;wp$TiM%>*17CfweIhtmE z60m%KW!I5%-x@N9#&9JVm!K~A+ruN6KjR1tz`g@ApSiRIh89PmN&1Es<+~lzy5WVq z7GT8&E36Fb1lFr82Y-~!n3=II6|V!CiRyUPN=c=$v^t6&Mlvh=S?s=9Q zuO1?t8U%;K46cfY8|D`J5hLvOtXwLcs;q&2r_Jjs38C%e@e(u&}J z_E$;1uyShfD-~TvR5l66$=~0qkXoR|{5N)dl0ahmo!qL>rLajWkz;8KH2VtGk+dUi z4Tzmve$m>;eKShUnwGqxdV&?SDy zXX+iAA_cxGeTx_+WNBD#H&NS{615(G7(kaQ7X@4bBs^gl2v6CbZvM(4XGUz_lz_JG z<8O@pMAK+BNhT^YP(ur@(Hp{^6lY{D#oY?64T6X2IB<~8q@yL>xhP;u1wDZy3lboo zt*u6}NTJY?&6%MXG8GzfTu8>u{HItU%Rt_qj=SB2T;-&$+y)U@GWWq(gu-OH&wP_I1=Rt_}$=) zFdy#?cA2ZL)@KICL{|zGuQ!9owCThwM|_Rlbx=eqmd5j;!6T}|L>ohec__t>uGxHWL@&idA!5K^;Hu;2`O+?+g!(>i4 z`vcYsT`KvVbPu)*?B&CM0Bd|jjDFfi-vp>ooMCkoalmNO93iGb14Y4NMUFohuPlZq z@W5Al`&*q2{@w+>@>9^xo)8uvS}7Qj6Nw&NCmMygar~V)rGKXFA_BC4uBbNb&?QIc z2l-@U4lY%~4^XNTW425t%Pe>D)L`k7k+gR+5PGB_)jxPx`g5y)uv>hwWU)U-x;5TGceA-dE;Gdkx<`4jE zT8gd%{IQL1bTwZ%L z33t{;rY8Jmydd;{g>OCj2)Zv=%L0GDk`FfpkG@aeMAcZ%!IFRguI2h0g|zNKT=K$} zk!MGeewv|&Ip?6J-rdgjWbFMNqufC%;AtHcV>DK5N`=)kHg5=S5+6KMVP-};ggU%2 zA5gowjrH%HUl}5VwMk0Sq0=CC3ZbQkV|%FN8K^YdSSxaWQb#xW;T)HotR*a&)*W5saQb@p!)tWf*c#4(G-?ITRwX~r5r zxuCYYyzGL1taYIEzFZIUL#7uS7^Fbl3a;U39YB|@bT^UMEMmvuAbQ=?W^_oBl2hl4?>wQNf*Qk4BBx1KMX zzQ0srM=B>|WF|NjOJI|&4v_RM6{u<{Jce+8vEPRpb@2|3BLU+(Nd=7r@f$#dVTfTK zmuNi~jti(mKyijnA^+%kl0;QPB?MY+%iNXyf2IY|QArWGCYOK>(L{0<=^`IsYT1W$ zJX%}xKGES1YkF;OKo;9*G`Zjm)qtOHTpxRPc(^Ou{jp45(>_2s-%#R^Op}!4!x`=B|1xIVkIhr0SFiE4!z* za))<3Z>o8F^q07B;aY^QRJaY0sJdYshXcpM|acIB`Y97sj10<7AO*R`ZZzfx*gef@faFCb`s=uVV;nlVFn zV(&6?AG4nJct(&UKSyQbGSD4=*H1~lNYN+ZHIcH2tzKF!T3aJYdza-LAsht|xEmc= zimd^~lYNG(#KZ%we``Q)wTG6!7corNhB{;7V9*0e#tiFfWx;?>24FCjOn4-tXXn%9 zwb^|hDyBbf)gYF_*V53?I!__UynZ>zpW&}YTEwJNRBm>OCBVK)Y_W5H1ePGoW-%BX zKT8zhM=F-;yBrcSZ;VBJgjrw<{QeHvwojhz59n}{;t1i3cuE*RAM%~t5I}^0+4mqI zPfn?9e6qNnZQMJ^rQhXWoy<>0cDQ3LHY_*s5oBPuFcPsaQ)}907|+z?&l`vNsXv;7 zX|EHy81T*zJ1SLW1@$3+Ec4r67JcdsKw&L^?_6TkbhCnMv`X@B66ag;!e8^Fo52kV!1nD8_09Odbgv^SQ)mf0R z*{jTQLayp6sQ4+^4++EaP5ZavtT{mAkIN+Q<07%uKTT)^jO0UqmUfghCd*T4z?^o8 z!!3oF1a9_{59D7IUhMyAf~9fFT!KVu5JU*-#23A^#Ke**#u|ku>>1pN-vHVkeqDn* z#$^L5vG^ar`KJ15wCxdDAb%uTtRUfnjQL-N2GljWqf#GiOLr3)ZGv)Amx!SIdg$BK zyJHXOB#Hk?CffLanue!3)?x&SEA7S?-%=Z}vg~F_H5ztyg2xCGNd;Mlf?A8rmJ4tw z5Dis3_}&cG-Uk3KhL#;5tvCOpg!|L2x7WbFRd@17q>;EZZU^V~L@MxI;X8l6;}smm zQ&|Bu7KfkBjLa8aU>aH7wM`1|=OBTBmU-xSx`^S@9bd42Nk(?pLIGE?C^f*;%Wc}G zO3Y?JQ@(hZ!bbwiK$=|*mJ?OyPW=sriU~_alro_MP&!R0+l!oVg{8 zV1y4I$pU22pnOgYr6urS1=Shabz+CG5>fztst%jxJsQqvmDo=IzEoIk-&d4szEyRW zopp`XAvMQ;0e;-Rz2Zetu_x@$z~N`7p+Pygkt+2Z#8`v#IG+@icSsw@Q`zVliq5eD|Iu?dDmnGOA$o!Qua1?60%vs7?rtIr z3O^@x#JF@ud6-u#ov(6i5Wi69P`_;8;-26*)nPZYEbD&_cAyNbp|(lXgzpU1Ae*AA zlQ2kst#TP9O@zzbg?`%d`P%@Cy{!;@p-4CdtDdxm0uuv~WAo$5|B2!%>}WKvtc_!6 z(`Erb2f+A|5Sa4I$ws589=#MP>*d~frzrWrY^L1;>;}5~1%))fn>D}oZ3?>gYaVx# zZkdT7jDL7jQ@ckHzRIB8ttBy`ft`0T7>1V6?bSjSvDp)1Rotkp@NDk|xZ}2oL%! zl@b8el`IWNb<;IE{kAfj>8?=RiBgT;)(WeRh}pj;;|$=dc<$u*C3!medYt3GZHTUa zt3a!ihW}VD`*L^B>TnlDgA{v5qVgQpw@3|Rotkj7_;bjmR<23FvaGKo0uQ6GeZ(k)h>z2MI;535ECJftshQ~Swo1tun5JLRo zV+7Mr=-^|P)CXsbVHC7n4%HHaY@dC90~LGlGoL%CopY)$JUMF`8T%dgG%PI_6g0wX zPL%Xx_ChSpkpMFC413%HdNgt# zRWX09QH4~Vf zFk^w04zPg>j4*5(_PJ{5`DMDkJqWZe?DX*}=9q{e1vsSO-g+kt)4Xp3!>UrX9_be350OY8LE5&z`uKg$Ozxj*hGK_vt9zr!tn#W^HnN^e@4mM8^hbuyv`X>|b zT9jMeWa(#zo`JII zq6=@x`M=2{#q*+Z)?*qjvl{FRvJXEX{0i>#|M& zjR-2m!N5Y=#*~=k7g6uH>E6B^?)&K=54Llm*X9(y#$&-dUUuYvfWz;u7@D*MeK+Wb zK}ygyAwy5#7oLzH2DAUv=Sv{X`yHz@oAbCVOio}$mhqAt8XGuvl!Cp1_h#_Dp2q@n zhR$_SxqzGT$;(#Pha&ViyuZUm@Z3Br5)V6vL>OiCcfwkvmYze-f4?zq4!r$KP*TCa z$FE%90$F2OLJEU_xh_R&$UA&nEb@Yh$_X>iI2~i}=<(-s6ee|5kraWvbD%k*Q7<k&@d59D`CP)P^^B~4UmSLKm)dflA-Cew>grT%RRL&Ra2_D-s2_gIs;jT3*OfoD)Z?D(Zlnx-X?1+=SH-%q?fU*k;k0MK?LWn z>hnDnYE)IG*Dm88wECvSQqzP=<66>GBq&F%Y0Ce9h?MCez9}2?f@IwYi%4~t8#R#Y z_tGtzo4~3QN8~vnynZ0R0TyI~*zfzy4!lLIxp&=Ko#0D%yot9)S2we7=8U(pXVZxI zv?<2+i8wXF-!5_R^?s0Q<6Zs(4Khnp4W1tSJuB6^p84hVG7N&ViSf)!Z);WU9hvO% zz;Tg(Em1=W^vOs1uQC#@;n&R3sdK+6e?jSv_4R3Nyb+yhdCSDi@S-V=-N{Y}%!32O z){*j!g>X#q3E2L9uO1Tp1<*6MyJ@URYWkAT+$daR&mn9%&QTMYTHIx;yDd{)T1FG$ z2$Dm4D8^<#fH5%yU2*HUl}tL2?ivmpBi*Tgv=omXuSA|fDi&o|OmW`@Q=@ere?=Nv zCZXXh=F#N~Fql`r*ELE;AWJa+Ks2 z`oW0fCzZmjq>zIIqU&-d%!~q-gs~)X+91y@abrH0jo1Qa;72RW`J%1`Kp(t?WD7(I zG-MH=7CUsfP?sTGArk^PHkZ-&E)%y#3NBJQ12H%_m(lkw6SuBJF8W6TF*%pe_bwH; zzl<&qYXdPkGPlw9E>~s)F*!7s(f2MBw?1Ak=L!QcIX0Kk_bwBcK3*>mw@PU*SP26$ zIXIWm_bwB+xN(d896ixFGgu{b95j$HZ}?`Ol59obZ8(nF*TQQOd%S&|XC{`lr6p;tZ(U_}1eX=B zgcOws7T>Z`!Qop~#s_@s$_1J?Rd5pDpb@bHB4rdhz7j?mbU^Nu+!6E;6c#BE2dB{q zmE;2RTA*WpF`{K91;_MkmmSsif32V$tz5;wq_>+7a5LLmO&rsA)sP(V-0u zb~YR%3OgG^BiP3(?4vN{XhVa297$)ej|VRTbU4B82n(7(XIWJ#NfKK_9U8HfQXCsv zhmOEL)>q1TfM7$VgTj;(l~IlWkmv+ONK{61u_M5LuF(RJI0Hxt8|Nyk9Cbovt;Jbh zR5p05!pBKE%bUt+!q__~fG~zOJjpGBsJwD~2N+l0dQzOJe4u?GTc8ypgepXQFd8h0 zY{4$)kVOz$97@FboO&KDHsO*^p%LZ5HVC;P(P+>TAYhn#C`TwP#a^_Gtbu?gRDd%i zq)|&M_4bwM4g4usI)W9qMi{eBLsxea=2SFWynTSUW^>$0T5L-zL#&{iP z!=7U6W-u9}j*feg=Z%MVqi!>fH=8*0qPgsK4u)T<=&MviyO?xdcScwnOYlhd(`a~i z8}CZ=7*9+7m?q$l=cCSGd`ru^Un}{5tWCSY z#qPtB;CuV(;~x}%i0t@=pyivP1r^YInjn5}aMAt$XRFVLo%a?cG4j!S6dGu8Pn0zc z%V26zDT7>FrM^+3R^Qa7eNyWsT)5R5j!Q=F!%T=$AiB}n zOw)oYZmhwAZ1t5y?c1WsV5f1*(Pof5iIQka`b49J28|XWwrX}tnoReQ#XZ9lTu!!5N5@H}F0BKO3TjH9p+D3#2=vamEzcv1k-DNEL71W}~bmVP}$+DY?y zEUwMFD?2xZv_H|&Zz){PjyDv(Y3sS5xtA6iUP2JZw8mYd)z4~~QM8mL=Wt|Im^Ou6 zXEYlQ-?d6@B1KvwHEN66F{;$6O^8iZ6*WtW(h4 z+5~-^_dMsk=l%4a`{~|u&-s17zkBYBJpE_z+dEjG#rHduNDGbH4~(tfvk7u$YiZif zw91+{DDQ|CWluGMis}0RgH=A;&gI6%?<%Z^CouvXp-w~J?|iLJ+RnqlmK~O2r@jA$ z2{Eem*U6s-ypG(Eg;dGU0diwZ7Xk_^0Kck?uS?cHRK2$2W!J$7u_4lwn8(MTs)!`> zSXlf4=42GIA#rD4iLC|3Wc+vxThz&yF>Y%TJJ+%Rn)RSdGC9n~1y}t(s`q}`9g^6Y zI`{Dtc5CYS1_2M~tVo4~sr=n!7somEVIth0W1xw<>iIq(cE&W8&(jX<;+pLEY6of{ zti@Fga=J-Pt)f8&*#~};N7-}UOMeWc@h2Jqz=y6#!*zBm>t&E8Q4^JmmjD6Dn^p>I#^ zhy2uXt+-QM9bV)E#^juHIPTHDwBY@GP>lRd%bJEp_)-(d=VQ@)tXRbl2KAzZK>FO# zA*qbN`_p2MB$mwQ!K2MtOnU1-kE#M6Xz;0Ps9qc%BeTCK^Ufh?-|X@-k+I;l;7Vjq zhDkczpQ@&!mWm^xi>df>(*>t^!_CVELoA%eWF}m-UPGy+Z1Y^7Ze-@u^~tr6k@e#y za#UG3><@2{qNi_ozJ}3a_%5cz5nhPRt5*o3qj}y8ptilRKcQ zC#$FZP~`v)<&^eZeo|86LIV}gKlTehmK7I%6F70NgD4_c1n<*-vIgY@9?^mwDGH^V zX-o#NzHNcj4fhtE{n%^%4!XD^=Mu(>hM*beVTtuEWV`^@wHV`tk>wJ01wm8=c zZ3;Op6vj;fS*N` zqor6q;^^8OLkKPv-Y~k59%Ra-Dbz&Gd7Z&;N;Lt*Oumj^)>_vq+Pd1p&m`tvOj>>!Qy|g(KN0OUC%+m%FtmbFL z{`qhuIj9uTZ2L=>dj#5o$;}@eP};;XA(IE`-SwwAcQh$4PWRp2$wP&mv&&^aB>*NP zsK$f5=+Sve93v~oj%M!B$^hxY8JWrOYnOfSs<_62!b9DaLcG5H+ot_j%^0e7x;)CE zA#7@&I${Ws2n<8Yp8wNN%g+bwhNOgtUz{Nr>>KD5)obaEA@cO=ZJ@e#rOqH5np!!U znn8ds5ErV$mtC|Qw9UZ>JZEJYa&CE<88|&Pk%{WMJZS}jAJlTqhvrS7aS}!hjXT)| zN-ca@sNneQw7Pg!h8JY6Um9(yo{1r-X|oi@gCMI{yvXk{y=ecIX+&7QZ`(SHUt0?2 z$}1ZXd26DpixCs3%srRc!~6^wI@K>0dd+yUF}~1oB@>paxeI)F8qiprJ&jmA zA#HI5I^D{+vA8#tqxxlwxw74JmsAvcX*r%$cHHUP*O38i6e`9WzL2YQx4dC#uqvKw zk!v(ohQw$@8^6_ndc6}j#Ncy&C?ualvwB6jB~b4-Tt`f;43eTqc3|sk8CLtHRCffkH}i^yMvO6by8FBBcIo+pZh+BQg|(z8~ONrvw0%reMH61Fy8G-2%3j5 zAQL(3@o;N99`)}hFvIh&ar?nTD5}IQsgm*NH|ntVUn{PJU_ZzD;srO9kJu#eM2e2% z!k>1x>>KB)JwumP#(Bvb4Knw6MH?7W8U)p>v9ym?sb2x@@Q3uAO>YahyMzsnD&mHx zOC#gtF1o>nbvH*+y#K3EjhwS!GEp79s9N6r7HyoR|^%$NkX$jJ^A1o;e2nT zQ2iW{yz=s`w^VKws9J-Er8&$!jU{1M{PStoUGblGTkxvG@4@Dq8PqvH3VG1ngw=@7 z2*bFlv^itHD*w48@3$rLxu55-w?58(s|z7;|5%Oax}PY9isXb*QoE#dTUdFGH}!ZHpi)72O3w=uAIG%omOt&^9V~%YU&}=_^1^EfR45qUH&s@2oV>&yXK1em zQ?_AdFOg1@d=BW%1xh4TICS1|#)c zAv<5Wfa_SeZ6z_P9VsYju6yjTB+*ptz?1CT)a{p;!FfnqNiXBZ#TpOi;nWjEguA4t zH}@$%Nliome|Lwz+xATiQtw4OLf*s?4QDKtoM=)4js=_vjztrDZb9~Vxw^s(vnrjh zZkK~&cBOqw$Q`5kqq)+8q;1lyqto_&Fvk4f_fmoQcUqMw+o3XVB|nI4HT0gLOI=N| z8_!M_|zb5Fn)-7Q6_#jg*ylWPrdtL*2l#vR8S8HQ4?x zSPp#mO0@9vLtKe>!7^YA2FOF-r+!yV_Pi~x+tHBp`tFCSH z8*E6#`+@NNF6p({3xA5x1u!lKBRn=A{>Md`s>L6e((zoGQY4r4+Yi>FM-(Pqpt;hX z^mZusn|&E3K^cZue&se*4V!*4w97J-Vo%UrV4w^=Oomy`EXKBhB2b2vD>AEx*`^Aw zxMb@3HkH^TL16M$JDaYa?b{JF*Ph?M>41!*or%aKu+%R=w+Ph89@u0JzV$e=K9@dt zL3$>N{F{QWQ*d%g`VuY`!esg=lb^{WJkC@v!>G}fp_+tNTd2b}kRU{}4=<<+L1Y&M0ds1Mh-l1D!B+|1$!qnNYJxO1cqPcF(H9=5mQ6h9jBne^EiYgKpJ##zRK9))RRT zbj}1`r?#rSe#6y~xKQguB&*r{DdD-Kgu_B@v5!_Dd`z