From ea15852ff3f85f9c5c7641c08cc61e4f0b339093 Mon Sep 17 00:00:00 2001
From: mhjensen <morten.hjorth-jensen@fys.uio.no>
Date: Mon, 27 Aug 2018 21:23:26 +0200
Subject: [PATCH] Typos in hw corrected

---
 doc/Projects/2018/hw1/html/hw1-bs.html        |  14 ++++++--------
 doc/Projects/2018/hw1/html/hw1.html           |  14 ++++++--------
 .../2018/hw1/ipynb/ipynb-hw1-src.tar.gz       | Bin 206 -> 206 bytes
 doc/Projects/2018/hw1/pdf/hw1.p.tex           |  15 ++++++---------
 doc/Projects/2018/hw1/pdf/hw1.pdf             | Bin 195801 -> 195815 bytes
 doc/Projects/2018/hw1/pdf/hw1.tex             |  15 ++++++---------
 doc/src/Projects/2018/Exercises/hw1.do.txt    |  15 ++++++---------
 7 files changed, 30 insertions(+), 43 deletions(-)

diff --git a/doc/Projects/2018/hw1/html/hw1-bs.html b/doc/Projects/2018/hw1/html/hw1-bs.html
index 0041f1940..829932ef6 100644
--- a/doc/Projects/2018/hw1/html/hw1-bs.html
+++ b/doc/Projects/2018/hw1/html/hw1-bs.html
@@ -128,7 +128,7 @@ MathJax.Hub.Config({
 <h2 id="___sec0" class="anchor">Exercise 1  </h2>
 
 <p>
-The first exercise here is of a mere technical art. We want you have installed 
+The first exercise here is of a mere technical art. We want you to have 
 
 <ul>
 <li> git as a version control software and to establish a user account on a provider like GitHub. Other providers like GitLab etc are equally fine. You can also use the University of Oslo <a href="https://www.uio.no/tjenester/it/maskin/filer/versjonskontroll/github.html" target="_self">GitHub facilities</a>.</li> 
@@ -141,9 +141,7 @@ IPython/Jupyter notebooks invaluable in your work.  You can run <b>R</b>
 codes in the Jupyter/IPython notebooks, with the immediate benefit of
 visualizing your data. You can also use compiled languages like C++,
 Rust, Fortran etc if you prefer. The focus in these lectures will be
-on Python, but we will provide many code examples for those of you who
-prefer R or compiled languages. You can integrate C++ codes and R in for example
-a Jupyter notebook.
+on Python.
 
 <p>
 If you have Python installed (we recommend Python3) and you feel
@@ -209,7 +207,7 @@ We recommend using <b>Anaconda</b>.
 <h2 id="___sec1" class="anchor">Exercise 2  </h2>
 
 <p>
-We will generate our own dataset for function \( y(x) \) where \( x \in [0,2] \) and defined by random numbers computed with the uniform distribution. The function \( y \) is a quadratic polynomial in \( x \) with added stochastic noise according to the normal distribution \( \cal {N}(0,1) \).
+We will generate our own dataset for a function \( y(x) \) where \( x \in [0,1] \) and defined by random numbers computed with the uniform distribution. The function \( y \) is a quadratic polynomial in \( x \) with added stochastic noise according to the normal distribution \( \cal {N}(0,1) \).
 The following simple Python instructions define our \( x \) and \( y \) values (with 100 data points).
 <p>
 
@@ -218,9 +216,9 @@ The following simple Python instructions define our \( x \) and \( y \) values (
 y <span style="color: #666666">=</span> <span style="color: #666666">5*</span>x<span style="color: #666666">*</span>x<span style="color: #666666">+0.1*</span>np<span style="color: #666666">.</span>random<span style="color: #666666">.</span>randn(<span style="color: #666666">100</span>,<span style="color: #666666">1</span>)
 </pre></div>
 <ol>
-<li> Write your own code (following the examples under the <a href="https://compphysics.github.io/MachineLearning/doc/pub/Regression/html/Regression-bs.html" target="_self">regression slides</a> for computing the parametrization of the data set fitting a second-order polynomial.</li> 
+<li> Write your own code (following the examples under the <a href="https://compphysics.github.io/MachineLearning/doc/pub/Regression/html/Regression-bs.html" target="_self">regression slides</a>) for computing the parametrization of the data set fitting a second-order polynomial.</li> 
 <li> Use thereafter <b>scikit-learn</b> (see again the examples in the regression slides) and compare with your own code.</li>   
-<li> Using scikit-learn, compute also the mean square error, a risk metric corresponding to the expected value of the squared (quadratic) error or loss defined as</li>
+<li> Using scikit-learn, compute also the mean square error, a risk metric corresponding to the expected value of the squared (quadratic) error defined as</li>
 </ol>
 
 $$ MSE(\hat{y},\hat{\tilde{y}}) = \frac{1}{n}
@@ -238,7 +236,7 @@ $$
 \bar{y} =  \frac{1}{n} \sum_{i=0}^{n - 1} y_i.
 $$
 
-<p>
+You can use the functionality included in scikit-learn. If you feel for it, you can use your own program and define functions which compute the above two functions. 
 Discuss the meaning of these results. Try also to vary the coefficient in front of the added stochastic noise term and discuss the quality of the fits.
 
 <h2 id="___sec2" class="anchor">Exercise 3, variance of the parameters \( \beta \) in linear regression  </h2>
diff --git a/doc/Projects/2018/hw1/html/hw1.html b/doc/Projects/2018/hw1/html/hw1.html
index 4111259da..adb5658c6 100644
--- a/doc/Projects/2018/hw1/html/hw1.html
+++ b/doc/Projects/2018/hw1/html/hw1.html
@@ -92,7 +92,7 @@ MathJax.Hub.Config({
 <h2 id="___sec0">Exercise 1  </h2>
 
 <p>
-The first exercise here is of a mere technical art. We want you have installed 
+The first exercise here is of a mere technical art. We want you to have 
 
 <ul>
 <li> git as a version control software and to establish a user account on a provider like GitHub. Other providers like GitLab etc are equally fine. You can also use the University of Oslo <a href="https://www.uio.no/tjenester/it/maskin/filer/versjonskontroll/github.html" target="_blank">GitHub facilities</a>.</li> 
@@ -105,9 +105,7 @@ IPython/Jupyter notebooks invaluable in your work.  You can run <b>R</b>
 codes in the Jupyter/IPython notebooks, with the immediate benefit of
 visualizing your data. You can also use compiled languages like C++,
 Rust, Fortran etc if you prefer. The focus in these lectures will be
-on Python, but we will provide many code examples for those of you who
-prefer R or compiled languages. You can integrate C++ codes and R in for example
-a Jupyter notebook.
+on Python.
 
 <p>
 If you have Python installed (we recommend Python3) and you feel
@@ -173,7 +171,7 @@ We recommend using <b>Anaconda</b>.
 <h2 id="___sec1">Exercise 2  </h2>
 
 <p>
-We will generate our own dataset for function \( y(x) \) where \( x \in [0,2] \) and defined by random numbers computed with the uniform distribution. The function \( y \) is a quadratic polynomial in \( x \) with added stochastic noise according to the normal distribution \( \cal {N}(0,1) \).
+We will generate our own dataset for a function \( y(x) \) where \( x \in [0,1] \) and defined by random numbers computed with the uniform distribution. The function \( y \) is a quadratic polynomial in \( x \) with added stochastic noise according to the normal distribution \( \cal {N}(0,1) \).
 The following simple Python instructions define our \( x \) and \( y \) values (with 100 data points).
 <p>
 
@@ -182,9 +180,9 @@ The following simple Python instructions define our \( x \) and \( y \) values (
 y <span style="color: #666666">=</span> <span style="color: #666666">5*</span>x<span style="color: #666666">*</span>x<span style="color: #666666">+0.1*</span>np<span style="color: #666666">.</span>random<span style="color: #666666">.</span>randn(<span style="color: #666666">100</span>,<span style="color: #666666">1</span>)
 </pre></div>
 <ol>
-<li> Write your own code (following the examples under the <a href="https://compphysics.github.io/MachineLearning/doc/pub/Regression/html/Regression-bs.html" target="_blank">regression slides</a> for computing the parametrization of the data set fitting a second-order polynomial.</li> 
+<li> Write your own code (following the examples under the <a href="https://compphysics.github.io/MachineLearning/doc/pub/Regression/html/Regression-bs.html" target="_blank">regression slides</a>) for computing the parametrization of the data set fitting a second-order polynomial.</li> 
 <li> Use thereafter <b>scikit-learn</b> (see again the examples in the regression slides) and compare with your own code.</li>   
-<li> Using scikit-learn, compute also the mean square error, a risk metric corresponding to the expected value of the squared (quadratic) error or loss defined as</li>
+<li> Using scikit-learn, compute also the mean square error, a risk metric corresponding to the expected value of the squared (quadratic) error defined as</li>
 </ol>
 
 $$ MSE(\hat{y},\hat{\tilde{y}}) = \frac{1}{n}
@@ -202,7 +200,7 @@ $$
 \bar{y} =  \frac{1}{n} \sum_{i=0}^{n - 1} y_i.
 $$
 
-<p>
+You can use the functionality included in scikit-learn. If you feel for it, you can use your own program and define functions which compute the above two functions. 
 Discuss the meaning of these results. Try also to vary the coefficient in front of the added stochastic noise term and discuss the quality of the fits.
 
 <h2 id="___sec2">Exercise 3, variance of the parameters \( \beta \) in linear regression  </h2>
diff --git a/doc/Projects/2018/hw1/ipynb/ipynb-hw1-src.tar.gz b/doc/Projects/2018/hw1/ipynb/ipynb-hw1-src.tar.gz
index a0d469c37377e1ddc39724a0496b8ef4e2121774..875d6394b11a266d6b49a5d3d37a17dae3f7eff9 100644
GIT binary patch
literal 206
zcmb2|=3sc~*AmUZ{Pz68EG9#NV}<AKj@lfYv?K6*Yt@T|3Xiy5%*ujIFRCp1{jF5=
z)yaEHru?g)-Y?mfm}XdASvx=UX$+hH;wiHvlxzLeA7`(+xOU&ADwSR9dc)G9`nRsw
zXTThI-umUc8Qs3qgKznZZ{p^Eu9y7!D}Qy~`V-~<1W#W}ubf@JwAax|Cr|XT+pWp#
z{_Nw{&Rlf5Hhz+|Ol?t;v--BonvaGz-WJS%<i2Tl3~&8!VFqOIBYqzP8^h5c1`P%V
E04nEP>Hq)$

literal 206
zcmb2|=3qE0))LLY{Pz6$EG9#N<Avw#j@lfYv}1Rb-iw3I3#aUJNJ%#T=b0n4^7prQ
zS}#uCTQa52zWwJ!10Lh|-P3=*UUXu)O_kKrD_d5joVdAV-;!A|_r;_9U8bil4UI3=
zb)NFo<I3l-UDnHTlJ<M;-e?#3f5ZCx==Jq8ifvu<e!ACe(L4ER?djYPkDRxy7T9*F
z?)4#AtNoMIZkYV>-V*OFzE<LXXQtHj*biyuc1J!)p0yKwe19Jc12XtG-8ei&!ID9P
GfdK%(1!8yr

diff --git a/doc/Projects/2018/hw1/pdf/hw1.p.tex b/doc/Projects/2018/hw1/pdf/hw1.p.tex
index 7baad6833..57890b7ac 100644
--- a/doc/Projects/2018/hw1/pdf/hw1.p.tex
+++ b/doc/Projects/2018/hw1/pdf/hw1.p.tex
@@ -164,7 +164,7 @@ Aug 27, 2018
 
 \subsection{Exercise 1}
 
-The first exercise here is of a mere technical art. We want you have installed 
+The first exercise here is of a mere technical art. We want you to have 
 \begin{itemize}
 \item git as a version control software and to establish a user account on a provider like GitHub. Other providers like GitLab etc are equally fine. You can also use the University of Oslo \href{{https://www.uio.no/tjenester/it/maskin/filer/versjonskontroll/github.html}}{GitHub facilities}. 
 
@@ -178,10 +178,7 @@ IPython/Jupyter notebooks invaluable in your work.  You can run \textbf{R}
 codes in the Jupyter/IPython notebooks, with the immediate benefit of
 visualizing your data. You can also use compiled languages like C++,
 Rust, Fortran etc if you prefer. The focus in these lectures will be
-on Python, but we will provide many code examples for those of you who
-prefer R or compiled languages. You can integrate C++ codes and R in for example
-a Jupyter notebook. 
-
+on Python.
 
 If you have Python installed (we recommend Python3) and you feel
 pretty familiar with installing different packages, we recommend that
@@ -246,7 +243,7 @@ We recommend using \textbf{Anaconda}.
 
 \subsection{Exercise 2}
 
-We will generate our own dataset for function $y(x)$ where $x \in [0,2]$ and defined by random numbers computed with the uniform distribution. The function $y$ is a quadratic polynomial in $x$ with added stochastic noise according to the normal distribution $\cal {N}(0,1)$.
+We will generate our own dataset for a function $y(x)$ where $x \in [0,1]$ and defined by random numbers computed with the uniform distribution. The function $y$ is a quadratic polynomial in $x$ with added stochastic noise according to the normal distribution $\cal {N}(0,1)$.
 The following simple Python instructions define our $x$ and $y$ values (with 100 data points).
 \bpycod
 x = np.random.rand(100,1)
@@ -254,11 +251,11 @@ y = 5*x*x+0.1*np.random.randn(100,1)
 \epycod
 
 \begin{enumerate}
-\item Write your own code (following the examples under the \href{{https://compphysics.github.io/MachineLearning/doc/pub/Regression/html/Regression-bs.html}}{regression slides} for computing the parametrization of the data set fitting a second-order polynomial. 
+\item Write your own code (following the examples under the \href{{https://compphysics.github.io/MachineLearning/doc/pub/Regression/html/Regression-bs.html}}{regression slides}) for computing the parametrization of the data set fitting a second-order polynomial. 
 
 \item Use thereafter \textbf{scikit-learn} (see again the examples in the regression slides) and compare with your own code.   
 
-\item Using scikit-learn, compute also the mean square error, a risk metric corresponding to the expected value of the squared (quadratic) error or loss defined as
+\item Using scikit-learn, compute also the mean square error, a risk metric corresponding to the expected value of the squared (quadratic) error defined as
 \end{enumerate}
 
 \noindent
@@ -274,7 +271,7 @@ where we have defined the mean value  of $\hat{y}$ as
 \[
 \bar{y} =  \frac{1}{n} \sum_{i=0}^{n - 1} y_i.
 \]
-
+You can use the functionality included in scikit-learn. If you feel for it, you can use your own program and define functions which compute the above two functions. 
 Discuss the meaning of these results. Try also to vary the coefficient in front of the added stochastic noise term and discuss the quality of the fits.
 
 
diff --git a/doc/Projects/2018/hw1/pdf/hw1.pdf b/doc/Projects/2018/hw1/pdf/hw1.pdf
index aaaed7d369e00f16b30f05fc73b5dea88cea8b58..832626724bf2527f925fb802a735f656651a9927 100644
GIT binary patch
delta 11743
zcmai)(?j0@qsFuCTDEQ5wr$(>&0NMW8_Tw}u(0f9+b!+9?>T?Kx&PcfH;>l;BA)z3
zOs0h3;!R70rv(yp4Lmlvaf0U>b`Q&!mLg)lu>?7^c%?eGU$0#|a!d;U#;~+dOwl*K
zv;~6ZhQKM+)$)jDVg|`VJcfLn6-(b~6aMte4?KJvJuly@m!Z#P@9Il+<jUF(O_jv3
z*JWc!dG;g<kKJt4`5DU+cWfV!r*tL$-52Q?$mhciRs^Jol<j;!arAC~SXxoDA12>w
zHmhf<mAGfd=jncw643K@C*RE}Y25^D9w?4!wNT#=R4<Fiy$*3l!I3B~bG3I>t{;u(
zTBh8-9zPeRI`l<!g`FrWqDe|Z!=>_RwONwSl}eT=#x2B~d28`FDz%_l2@cN7|GbJS
z00c?VB>{scT*<#JO)9b+MlP@~?)?YV4F(;WqXQk=58_yEQom*mT)qsA?;XuM44YDM
z)ySbSBdOhhpA<5jp_)w-DWLQwS-()iQ)fPgHXKVos_dy2+DWXmg~GcF+{HRa7i!#q
zVLot3yWMF->b`tu5~RDgJo)7(NTumJ+AN{0Ss>Z+n!UUOQILV$WhQLgY{pdk$Fzin
zj@A{(3(|m6zUp-cvC|x2%wmJp-8G-{PCD>@JZUWmh+Lxi;HadfrI~V@v82y)$L|KS
zf8jg&J6Vt|#v4u)<BGCrW>QZ9Z9AoL3u3w)faRn_;5z+GBUYj14=Z<yX$%8Tj;gPV
z1%7wrg40D{w<c^RDu+O7Du<@pb_V~rw%-&{<o)4$xtDNe3qd>*IwmV<YqV=v9BhNF
zmUpC|O5RgH)~Di~U^7J?@9=(_UbJo?(EIIXH<i8qpk^;r`{)jZ<+-zs#@;D3aI4_p
z8FiV6$SMjOJ%1(fey6GWLP!nWrm=_ZI{;NE%QVHPUSjEYHuDdCUfdT96H*gS$!iBg
zQo7YR#0ZT2uRD5Rmur9EV6A)qC7T=bgjC^{Jv-dhmL`M2=1@*>6ct)>AJjydWw$EW
zL~7)ZM<Ky|*<r+v*C5D=AAN{IOsz2S8{Uu1uj9eKwtZ6G86I8Z=LM2bMpGq7nLq{_
zaf@tb9yU12ZdBYBYkVQo-Ne4rldLfe2Qr%}tI~~=9OBPEs1up%k}=a=aKiBHFGg+K
z>Z-8%MqDoxvL`y-nqZIuuR;bsD*1IQCdK$=b1melev@$_h~vVTc;Ox{_y*ya_~fh4
z^^h9b#J9C%mXP0}hTx3-z04wHJ%QB0S0TIR52vlY?Dru)uwL7ZAuyznm4$y6DlCp3
zmc~gk$D%Lvu3YnCl~$B3c^OkCL&O~&6wRTM`@>L_R=CuQv7@(%A*kWbj)B}+6??m4
z;$p+F1su$)HqC>YoZ!o1IcI+>u&cVch3?v}qcm>JMP0S&^D;7TQ1L==hXDPtd3{`C
zF#*Bfa4PV4LVcFWFb+60k=xEQ8`|Lq*<+ryc1XsXqMXr?zQXIo<ljF@_K@KM6lec*
z>h(sYj^ejVj@Y}bG2(F}qC-nZ$AmR(Ud^spE46^_m4;7hoMo6-VNv!WHvDm$dYwrU
zG5&qJLG>1VyQn~;!sCgxB@d7YnVIrD<{ge=40o<)p?J+>sPk07)ygmUF7iLr((*?z
zc>KKXLu%7Qwy^mK+kLWPWmMX*Ce8@ASFI^iHRoGC9bN}NhSkx6{$)m8jB+r%x%U~>
zc(A9OCLv!51(9w+DktWR`;xb3_Qtn-_F!pevuj?|XrSUFuR9j28VT@V{<`xGZ)fLH
z@+O*v0TsQ#3f*gzFNWPsGxSxpCqvRl|1b};W~r4iMsHx+)es-Ta_7P@3Y?`AHhIMi
zlV&ZjKO`)>e@~PSj}<BlkHyS4#wt@i1?#=yaYC+|ja3Ns&EZ-SbWlxgTJ8d)C~jU5
zEGy67&PSwAK0GAz(gJ9*F%YzeG|wG8<>Frj0}5))B<uCme7%yJJj}s5YI1(#@%mH{
zcL)?-rYtouLc8bTB(FyqZKEXPeZUJ(mBMz3$F&P?kX1E(yTkJB*&&_-{|D)s;@mPk
z7Mml+L3vMV-}uoXo*#l>936OEI*2%we}Pr4&$r{<LtETw{|2-}PM_iQRG)C@aO^Rw
zpW><uiyT8hhKNgM#>l431tL`v%a+4pq#800f)6);(}7GM4k1rbyUb?7FG?>$hq}aN
zVA)cxQ?;>o$Z^qEanFqOd|L5zI@ppQz1a<WJ48UA^;oqxXArNWMx(Byp6<ncAnz~k
z`6ud=bVD0AxC`Li{KWseU{Yn{^)%Afq`NXJ7Knia%`2^a7~Ooa_PM6I5n3g5H9HQU
zgM5WA8Uo;UBL#B@{(EjSmnY8ahzi_LJ5}yH*3F_`(p`90&~fuwG$jk-*vK|As+;D^
z#`0ABrxK`YmeBIC3JVY3Q9}(o$w33(>yM7p1SzS^GzTcm+$Xn$MR!^Hc|G7F7ZrFh
zAtMz^v}?;ehQBu$t<D`FJiuZ{o+l5$$=Vk*nZB*YHQ#L*q7Ah*k9)Dab}b~^7DME~
z1FP9Cf1k`u3Yu!CPu|zov4#*Q<B|)2q0%6oqrGhn?=>ha&Bs%vwmPL7j>y$A0>=sJ
zXcdTyHbBb$^s@?`nL2^B{->fupO&dbzviZqlX9;dYw(-Z38zz_yk|~oLPg2nSzCgW
zqmCTJkGx?HzmPc6AEFm;p;I4sh)P7JyXsm~5|3c?-XryCp=_K`2=AN+@rW{>?V%?d
z;<XWD&(4o`_ZP~6n7Xjxo$widpuHU&B-l<24S_>6T&Y*zrc&`ho|TO8t3n@&AZL8L
z>%+TY-yg*B)LE*bii!6Swp6Q%F&fu4BZHFGA>bOjS1m@%(?r`$)v60aMIEN8pIg~^
zz+3k*;~%!ih@UTIta`>%f?>0;Be=oYzw{7>xX{sBGX7b1lQU`Z>w(~CW1xbSF~?Dr
zasZrJltVl-)RYvERrW5AP%GEqVcS1#gptV(b)x=^;X$fKc1l)KLX52VU1*=6FTpUN
zSFU158QhsQ@N?1OW!?kNTZ9QTH#g_l%H*(NCGy<?(J618`xezpT?6^oO#gVFE6nAJ
z!e;FZXtXJU9V$J`gL6hGK-UzA*>UQ4l6JUMaYr}#_M_bYKnwrd+auHnkIQWp+2TP3
zBt+WI<}otNA7?A0$B^>hYV+mdd+MTIQr|fWr!NuicLXgz{NK-A2BQ~{(k$U=AbGjC
z|JPFW>1`%%PM`-BHvBf=O<exfe+OYff-%@58v>Q+g9C3G0It6;8$sosnSPlqc-e8H
zsgq{zK|ABfx`vx_>7eygRjpCy+oU%dPtveI`Ov_lv-3%sCdGU&Nfn>F8Of=A_>&nU
z7oq%W-@5M=P0sD}?D%G+b}YI);oyfreeMNNzQRv`jq*#$Z#Sv>8fDB`hJ?DiM_cdF
zb&b32_)k((LT={He75>Y+xMKD@twpsRyT!#=n8G}F-bXf#<H05P{KQL=CT!XrDmy;
zy5H;uTbE>4JI+l|GZja<6hQ4ZHNbzn%5zR{|MtWb9aZSFs=%4a1hxKN;Iy9cvGEA>
zjTPIUCw}X>uNRewMfOJB(p+)~{wpt#-nO{E_NGPurvpEeWT{!BEggqnt#+mTv^(Db
zLVx?drFUSH%Sp@S)noJb)Xr<^PLT&}c+>%r$ZEMxsWIrOnqI*9S-f)I$ZgHL0an5)
z1=Pc&GZU=h|9nnVA7H@UhhkqAmJ$z0%A4({bG>TUDcF=yCM6nm9d~#w6Kx!vSh%Qj
z25o9H=zIMttOZvhM_%AH_bo1F@;O>{aFDT)W8jCDbH?!cjV^@5QhPgVcF(0C)YLuW
zVOHDWZ*(bl%=R(xE_M_;K4>;=UR5>$A9lz}Vv!BwVH>TRCDo(%*l+{ny%7U_c8}du
zU+TCu6~?%jbMd;zx`%qUb`Ev2jf4gp_^cF=^goQeJa^gjCW8bBPaH|E&vGn!=(|>l
zWXE8K`=EHv$!fb{xJh{Uw^R=TzWN<<l?*OlP%@VK8O~>#)^?GsS@)5gTV*jOOCHeI
zp%yokx-PXNA%zH5#)@}&9QObeUnv4lB5(3eYB(OJu6ph0a|cudg-&K~RM<60iBHxj
zew=eKWNGCmwiWBKx{$v_t(95<b9wT$<+2tomm}@2#VXQLZ8xtod?W=%Z3CNTC+E+h
z->wm&FIgI4c`nIzq8`^kkWqGYG~~uBORy|HP3|`qv7N>QrKLWPY(_w`my5uOa?5LG
zmuwx6&t3@z1oN+^<D@xfo?xHoYjO8I*5w<^k91>d;w2%^Xm^$OVWmU3EqQD5MLX0R
z@J&n-)vn}f5=!R3+v;M4y63nry&bi4?R?Jd;nvjTVAx(Gatx31q?EHX+T3&)bmZMF
z?nh=~ugmtXBtHG29cn=QTqyl{<AFs;;t-qZX2OkIyxlivzg}LYe8@;SPRCs~u^Mm2
zmglnzhh=UrV5wo9Jo2J5Y8-31{<Zgk_wJq|yQAFY9^&FO#`lsVaBoJQuiPsKr-1Xa
zT1zkOzT8wVC_;*PvaXWyU;nJ<q0URJ-C8)rFcQRrq;B4uVjz%TG&Cq;$SfR+rZA-M
z3>*#8VA5IIgZFD%X2b4mzBe0bB{G$9t8*x*>~j++IZ9OrMn!0^*kK!3Du9mmDL2x+
z&hVlfWXg`MBcRHrazto(-3vz1guz>JzTR~k2%+*sslVB<uq3SDW3(ZS`Q2hZ_QK@s
zbMuGuN!~JFSpe#xEl3qxB25X)5I_A}6M7v@oaPy67W<ukwf9mQM;$1h93hsV)@Do)
zlb<(3I5$--kko^?yDjM`#J)mmO;>Vhyi?IkltrmTjJVU)vVV6W&@_!h2wbE>s2J7H
zS-IpNWczn}$`f+iH%Ygg_~4Et2E2V<@_CL65w`N)p#i6;2Eh+@G-4C3f7AZ*_6I3S
zlp;E$BVb~0-`?7oZI>am%}Cxy<5SGo<;v;u3PpNhGPrZRDnDK{>1x<D?rU4G&s-Nf
zKQp(_XPu{-(l;@4z~*k5;3$&K>9)dntbJE*<<b(iG=4je0g2LDuS%U*K_fdFuDNq)
z*Hjn@JOk1~oZ&g+Uh>>r@~QE)4I1`&>G6k}UI`C#7BmEp75wAMD}PUrmd0}_KBF;}
zQ*44U^E){)rn5anv2rV3E_pZ*3BZ%iU?(ceTBS#;{3xsPM3(1!r-Ps$;4b(dbff)T
za*5C-mQ4Ghrn6k%u0yRmOE77~gj)zQ>ib*@y#S}&xZZ|8v^ROt^{w)w-0+ox;6_u$
zrgVFRDyp+ikm)tkQ5psA+k!ak1-6hI5D+Z_sUFE(j3!F2uNGRA2U)rA%@(e<CIdd6
zalGjG42VkGYd3zTK{ULtU9ffW%f<>mm(GI{B7uo<cHLoT7b@@PU(26`#e+4j^ZbTf
zWk4yy5+$<V-6KLY?S*SRL@m`WyFPf$29hS!bT$8L^?s=7&m~9`Bz_e-9YT$jU`lvy
zXlgA<gB_DUxKL`iUk8|E=|K^EBWt)MTSe_#t^Ma0bWR;LD~5)N7l*t3w7K>Om}7Rr
z^}NeJt`Af)@mM?uT3FC0z6^ZH6QujY@PUMT-cYKK1O<!~A0lN$oR3p^@&W~@^=4%7
z-wL<~y&2YLuk;oZ<Y^X$;we~EX!6~Le)hH07C-YAZ;IG2li%N)3k7sud*mCK;x{GM
zXB$s6Y7>W&H9MisPg$1Dx)|cy`xUu0<Ji%r>9UlQ&*WL8n&_(i(PE63;5e29&wv#q
z1bRF?tfkQ+-cVEf5K7FcQ?^yHBT+Csg#CiJEU<U~??E2JUmUYiBHm(L*uN|*1EvcF
zl!xIoE#6CB+1CeiI|tUhkO%m3fB)()+a%_}?=ADH`;oQLZnuqqlDM?3y$uRRVD0x$
zNAh1!7V+9wf{6o*e-F_OXixTQPX_2%q^18{vwYJ@JWwMsS-VQ40#Yfustk6({Hm6m
z-|r$4L-pN|ycp8xAEM)VxefUXUKw=qw-vp`)F-{dLbA!kXoc$lRm-{3Lea&R(<{YH
zot12g>l-3b;Y>N=Pxh96Z>~s^GiFhn*xv6$VMvaJHc{hW!=RwS`KqgqYydg4Nw#?&
zC26%iq=x{W;Y_nx7;=cF&X`g%P`UQ_{Ay;oK}W(0s}vzEnOO#B=E}Tn;O&;qxOQG{
zHHD!6V+MKKg5~78XG}%s-~IrexWglZm8W5<&!wFk3f(={;nuP)zA`j(Su-adbX*{T
z)vnobnckL0w<RYvx_mk26kt7mJfs)TS?W&Qd?I))!a;6^H7%R1B_~pwv#t^vck&s$
zGwxx-g#N`SwD2W=pm?OX+q76Z0|{w~znut|wkcW`(pDn>+cjmeD5Z*;RB_r>!5zo(
z?O!VG8hjqflRT-KN<5@R^9!D-bo7xoPO3X=!F&<VgCA|-3N!nVGQd9#)qX;LqH9Zl
z2pxqp9V9~Gk6<wDB}wi?`sP=~&X6i-*eTIS6lc)gZ1UvQFnKyxpUpvdE{7+XlGoMv
z>Gk9iq^LGr;g4lgfUQw6?s$^;x7MFNvV+?0hsiQD_CMG7ogMgF<kDZbOBXS#X)<LG
zf!39J1G1Njj}z1#BY?yZth%e7Ot-Sw6lzlAtyZ=-pf!1Fp0G%PS#D8ay{ENx(*clU
zAd#S+Q{%o@S{5KrZXw$y1&<~-o1ki3Op<?J$hPON_*V||kpF=3lnyF`2n^yQSgN_p
zgF@xcL{y?ePn9b?oQ*d(4CW7Li`IoCF-nx~GkvCMAhdb=18B_LJT2K+0WZlM&fo}<
z2v1EGYjzkd_}g<1y7LKvJmDI))>QZ;-is`S6XAWC%G6!D6|+<<)~A>5Q9sd=htSD<
zL!q1s(u!tCvWsRsFk1bj!&W|C9(FmC&Pt<T6(nW+vW2?`?ZhGQ2bIT@cNX+vLfHWb
z-^rE`8xt5`!E?RAU;lh}`*=A%@M;bt&F2M67_;eTW;OaXWi4d=Gm>a`PH*;a#aR>c
z#Xy=r&jXQ`f?>ekMWXc&fwToo8c1&T{~BpJK-VBCR~j$qrC}Fc;m-1Uy7gRP*^odh
zBektsdtH;9yGOds43alD=*;CN8xgO!Hcd7NB5K4Z_r@J2XN#KCudV3r?ZN-<gK9CB
zO4}p9)&+waDur>Lp2F~lWuMZY%9oxJx=aQ#N&kUE>Zcd>?W60xheq-RCpV?Y+Bhd)
z;L;o2kM^FPG02!c|M=Y^P;G}%q;^LTpc`9*8iCg#gz>)uWMc5ew9{=V0!Af76+c)d
z?dBoxy`^X5>B+y#X@_A@IEor;)E6$$%swiW;fq&Yfs62o4)4N9(8{eS*f<B>nfJZ6
znhq{RP;FU9=oNL6{XDhw(Cz&vn#N=d^vltruTZ`a`^#btK`K2@F7G?7b;TxxQ(Er{
zpc2&BR!8D<V^U|?H@1knr75Y#C8(zA*84}-)0L~G5}R@E30|-<B4m$dsiop))bcte
ziM@LKq%5L@jUzKh*PcB+7;V%@KAV;%JC+{elV&fk=&_=i#kawxSg5Y*ZekS!pnj#I
zpGId(T14D;g6h2+m#65>J7rg31MgQCZd--Px0zIU6f)oB2S1y<xJ>nGF}Ah{Ym{_W
zwE9_ksOC&NO_*zDzB-?{jtYm{w$uNmZ`+$3DoJHB%s9PWDEU6^3Zu}IwbDM{)jOBO
zT(7%RRpEyS>%m)@AIelHm3$`!V03OKettvY8gdoeK7lFZG3)XkzNR>?uX4%#Ky&qL
zL#_!5fK%z~aq{im$@oU&iYvUa6y0NiTTCX|IVr-dqip9WsJ8*x>+X{D5{2RL)Ww5O
zk$JZ}Q$*O5rd~Oaes#RoUd%AoUmmzaU$a@P^H-|Z$>QBU?z!{EAOE5p7+R1JJ$i3X
zTsUvU&^P%Pe43EE(Q_zX%44O*N7LO%b%H``nLiZ<Y32BXM~O%TAD^$)IOTZLX(w@$
z`3Hj8>?R8M3rltPrO<iYF`7v2zRjkFt+cxc(YRe)q)g7XanIUo>fX|FnU?KE{g&D}
zzWncqX8*f#Tkbzl&YtwiKz>-aS~1_g*6}iRzZ-pLY?B9~4_|6+$V}6~H{q3GzA3eH
z4O9#FS-fNa_upfNhWV#Kdd00_{S}>K${n(x?6=Bd1{Ll&_yy3+@+{4T5TzA!%l+*3
zldz>QQ*%nL?wxZ1>3vG>Vj^w^;_DDw*Z7nQvxrsmAYKukz(GSG0Qd7PTGuj7_tj>t
z4biR*8F5up%`!_WYP;@jh21ViSQ3kf;9-m*!iLQ1W92}fw1>ZeoHDh<R<g5^Y<Y{&
z3xh62&Lj7+<HndN^@TXx9E)|OC5=B0xN`50;s3K-48l^v#aTHT_txS5@fzc+!TCf+
z{osI6&(gPs5!AZ|5Z|A#mx_;}$Xe``BKOz5yZbh=pfq)F+SM6*xvpqg^rIpUbX6qw
zV_tQ3JKTogL03Y0JoIeBmVICLGl;6dg9ObTBT5yhCLk<HK93$I`~0UxJKIAw>}2T0
zPc0Qih~ULezo)<(jWCGa`%dN=+H98CjSh(CH+z&wVpgmJvalY0RuGl^rMO7nW$<;T
zHM2tp=k@%pZ8K{J5ssiSiTozEq}U!f^~A>DkP(JhS%qwcO8>K$gnQ^e9qzqzR#&mo
zFLmWQMzw~L`!`FqeK_xPvmxb%IEnYx5>djlF~AFOqQDfUP{&FN)k^;$vDsKaxQdrO
zA_Tj;N_nvYI&zgP#_Nfq(@dF7xaJLLD{&GjE|mrNb!o>5iESXEz_5&kXLh+@;7$v$
z66tajkc>-2V&Vz=L1TPodfx|c-0atz-oL}~%goJCI$fS2dGb&|8+!;*Nan%sfp~>s
z_yWyk>D~d3K<>F)vNg;=zSz0!<;q3fol4@SfV3*$fnxhFnOKxWSe~w6Y%#ffqa~C-
zd19jVMDg+0=ZeHthaTs=Cx82d8fFmT1zJe<uy2Xy(PDXkH=Mf!8swkd$BSswe~@B)
z<XCz=&6Ol@Bl3URWxg98PecTu1Ws-vix9isWLu$uqB9C+qoqW2GXz{hYOp}OY4cw!
z<D+X9f5lT$fpS%8w&cWy9~@nm?pu@?-?oG9V}EVAI)>|vfj(S2ZXme=#L-NNm0cc3
zNHU&LwSs5^?A5eAW{ivUqn19{Eh@=3rtTD}HQ0x_8c8upua`0^$7A7PgdRlJc1#4q
zv%W=VeFN{v`UbHwO~=m0&dSQoz{*a;%1T3rz@p@0DQV_uMM^8l&(6xm%gW2n^M42k
zbP6#9&hy{82NZe+K?v>*Dma6{1rm({+NYvKi{L$hAZV${5_->KD>DQ;2{S||r9A|P
zATo))HU@L_awbbHNrF;iEo_YKsZ%2SJAA`Xg(~yn+p1cobmlS?tzuiPE5yjnz(|#3
znInV*?2`D_hHA=Zp4Zl3nFDRTNr+8)J6EwnG8}0OIYeqP7uQ1gEcSX*Ahk#{v>ss=
zN7^i&9tTHU%lKMXs<4wgS)qE^+~FHp!8lF^9cj~ZKRu&U=!}t8mTlnPe?)Cs?Bp^f
zEK{o@k}=8|*Nmxl?Z1U_4WS|#Yz0Xr6X15k!Myt!K;eoW#-lZ5(XfK+BujB>t-*K4
zgx14r!>d9boU@4!`V)x)Ny^{bV{%wwc2Ow6U=20K!*L<a@nunJAw@?Lp_{~UnZD8y
zNkzv+D;JUzS%+JU&GKR(eYx@GvD3@lK$*J;&GL2N0~LvN)kJ5%-JncKi7dE`L)&7#
zNV!;HM5M)Ze@GKRI>|^7$s&lAx&0PsB~73qZDeQF7D(}hoRTU7HcQn620pc8M#Kdj
zcG@F-)dN>pSHi_FdA33OB>x5<6MX~Nsu1r^cFX>t2V8a5jBnS%Np?`6oa+S$6RmJ$
z(V+HLC*=$sJlouL4Tn%bOFdb^cBr_Kxc##1IdF^D5+0$k75P<npnbnE2FlO4`Wr=b
z*O^betYO^qMqUsP;F_c2ik=|ev0ui`p|#c)eOohzhm77nUw~Rm@hta5_l!O<5vZjz
z=00#Sf9crqg>4eJ%<*PK88d-pIsG?ex2p7|5<?;qoPMHYkZ8bmF`Ja*Pk-LRw$UQS
zL+rW#YEJzuLTqnWXZus;X-?+pwM_78RvMLTuwk;OjUKrR$YFr|7}l(BP5l@LvBq!4
zvKnL*zuO4a+Yr6F@W-&euY4bK;~0K=@#lZH5hw@>kE^`0rBu$je*Cx~l47f%C)!fK
zqp`dzraaNm&RmHysIl<gP@eME=P8|+;&YrEwcH@iM-@=xyDF)suhEP-QhUE(R4BR;
z8qPx~mnowFGQ&eL&viHCKV@`{{r97j3K5Rs_C|HuYYo|iNfP=UOO_PR@>r;PgAmmb
zW!z$54U#c(u~yniGl-F1xfX)5w8hYE3L#3Ti=)DFb6)7^kA)+}I7noA-MdsK>!*Su
zG%{9(aQ5fy1O#h~#UpBH-Xnbt{#~v>PU;hCfFQB~_W_(q&3yx^%#1z*tQJ?VscPb>
zD%!{kb#kaRQw5@#D~UlhfxL|HJJYqqBpAm?AJ^n*LJV2_Gy~CWoS$bUZ*8Ygk}8`I
zX^YNdVo~940$rVul`!o_`10*qb}6<A9py60m0$r5I@HEP+m`PBYz`sTiAhQf#>i2X
zGo8VJ9%{vbb)dLu{N!I9RyX!K8xH-r65E@%`PFrQ(?}A!fwsAZRO6dRoq2);&RE}G
z<T~j;Y(Gldi?tfz(`7d2`L*n+c*bIsz6LN@{N13iYYLJy__&fNXtzE81QmLc93L1;
z9l6y{U!%!<*Gr<1{xp|LKV<V^2v_|D;#b`U2;`wtPW)D<5V%W#83mg(y6K*R=ClIE
z(be`(>Jrop6yc)5$K~x7X8JV@*hEWD@7aGqhzVZRy6{O(4SSWFRV3z5oabKi!l0hM
zjJo;_5mAn>)B!`6xxzXvs^6cEr7Ui~2RJed>c=beCQ*LhJgXXPg;3-_0m24FyJHeS
zcxvIK9RNu{#p$$$-cMjNp{S2ae!;0{y8l_gNBcM1wXpL$YB0cY-ukWg*&Ou`Tbo!K
zq56H~T^v~Jv^_y8caWiv)*Ujs?0T=HzL!Qk1+lE088gI|i9_3hgVQaqcHmFHo>BO-
z7Kvr>UosO#4;^H&^eWxo8{tV0)OsEOPND|15gmOxy@bCbWm1-uS2}UJqEO9^oDvjK
zqz{%hwq}?p<=$d?25bbs^`#Cr<n+5za}?pn$)w;kC%5oY!Wyu^PuE3H`OQa?o#Xv<
zpY$)2)u{HDE;Q`g!<X*wKlv)O<LlZ#bwHiB9>&en9^h-a)uMt_O@wlL5VLAPBt-)g
z+6WwqKgPjCmZOStGz6zS=;Q%!Cs#-1&&&#F)x@_!-_fPU@mygrll&#p^v<`bHo@4W
z<(-+1J;Y9vX?3Eyv1Tb&+~ZaBNa2Di^3ggC>{GR^6+SpKbKa(7Z?kONJ?bD0oZ`$;
zY5k{KZr&(H&WeCOzZ7E@rj;g;R>dTQBHr6^8_fwLpXWg)*(wKXIh1Klq+6CjG`Ao-
zc&C2<xsk%1PgSFm^rfAvAo{PQkBP`)y``Hx*USp`-d>Iy8;o4Y++v3-r&-LEFnj_P
znLgM*SPl&q6CS_4_LLzmj<?!0pE9l3KDus}orp30H|3U*YJ47s9%cjpr>cZQ40q8F
zTj3oWXLixv3CpjR(kt^2r+EQs$2};TM{v_Z>=oC-dL{{L@jY~e#J7%KKRjq^!42_w
zPqyJViO^TsMEd7+3v=3w+%IkNb^oG(wX9lF!7<yBUr>4LXwP`vOQ>-;+N>D<f8>+P
zxECL-xxPuASFZQ8+T||+6-IA$b#R$WFZ0cMwCX~M*wmH!_nfpe=Xqa_6-%|u4vg6m
z<oNle>u^8JD<&Bp5V}Yxx1}17jPColrs$_Kz{lHHU@2ZSuq<-0fB5d@Wd#J)>5dEK
zAbAt`Xq}yqQZ2J76$;Pey&}-DGB2$aDnXPBcd)_%;XH4Drk|Vuk=pJ3b|##o<Ma8y
z!47wJ`L!k#Msn4GUx^kPR?<gKD7RbGamH~sRBw5kseaONq_h0bF5_AVV;gzYmPJ~x
zRZ_C>VFsWTx43>c+OzkJYV&|P^>S=~zu3)Sv38{9jSDNBcYuzT+2L`Zw@4Qu*`;+E
z@DKyHPz@E&<HiC;)kQ)*LELp|`Xm(5Lskx_DLa9q9+q<J>#A*6XEhuJ9$DejTFw3?
zi^lH3>47{}?qglrGplkhPd=@AaIPlCLi29%r3fU50TPM0HzUdKa@mS45EwYzlYCt=
z&}o?!Jh^-2Y!cCto<ul~OmvIARam&Lw5#MKY&aZpAxpqX<}6nrWo%|3Y1_);Y7;Fr
z1l8@4?s=rFRjhP~j62V$lYMN7NF|?mnedF$d*RJ7cipNB6Rxpegu((T<cZOJH&`N_
z1P!pL8GP47r*4-Xn35o%Z5Z6#r6D~An*!w>TzzhN4NI?&W8<po$}0O!Av?@tKg?jx
zrlx$l3k2+2&=);Se3%w0_jWZ1>hKWeeFkoRm!Z#MC~Qb$z7Imw&*3RJ<jyo$nRaI`
zmK6Gsc+pSw&EYFA@bYnbh}*JeV!9;PaTms(n5;W1hF-j9%Z_spPwAxo5t|dQ6+ER<
z4%^l8P=(>H_!BC<tB4P`^X=HMV!a2wC`h%IbOi8hZAzDQsLIvH{RngDk)$}A89mC|
zEA5xBr2R^FJd*Yh8~Is+Lt8fgW1ZTT@*X|*#^85_pGBI;Jdw5Bm<K%BHE<>*RTgcl
za=z}mq1f%6LtNcQoM(1CmZm}ih<xmT8c;S`hIF-)h&xIt(7waav<y>Djek5op$GAN
z$pZ5o2bS3C_WlUn0^b%ibfJcbns~AV8*<YY1LeLgY*FAh{f>8LwU6pKNK-S?6xeU9
znfo)Z03on1(GO1t;gE+?@%ZB^bPZc&wEtD27pq}3lo!ioOok23WK5MXR-ce(Xa4)B
zTYxY&&Aw)yv7+)tlkF%E8-~SvoVQ_oY!(ogos|I7qA?$P?Ww=IAKmGjb-3X&7_VSr
zS|{kajkMY#A}k#I;tvk$MqyTQomaqE;V|rn)nR3Doj>As)pDaSR2lOa<S28|orQry
zD^1<yHGs54k$Ne`-T|MO6DIBNMkY-X34Xi%@&UWI|D7-VxxQ@i<`3iTsalwq1OqJA
zG^s@~P8PSbWi~hKi(6C+vumN^60-8_?!PtKu18ym?_hjCKe%@DNM){&l-6nSAeB-X
zNXl8a<m}W%@qC(dhQ{#vCi{o57ifFomPmc=gGa3kk@3I(6=}TFw8AK5VndfSQOHic
zEUj2DTx#5+(uYYhQF-p1d=<qQRSQ^=Q$MxZB~2LSU@Cg<pH%PodVEtXASNah)Ekl9
z4vEdPDA^J+Gw0caH)|v>=t=PHx!D=mdbf}g1o}KZ_9&O*UdZ!wzC{Kg)+Oqc^=SRx
z1imhp50d%_-RdkKnX&V70}bv4d>GtZxRfHU7Dl<g`WT4gg)HMLC$Q&U<AJ6fK8s#-
z;JEe<<crEsxy4rW;+?b3Lz&XBICN{Ut!A3_!pE!%=2S8A&~l*m$?*<V{_T&!5>SYm
z>^{o&$rbi5If}Wwki(4|PkGag<&t?{l?84l$dR4DFm_;U=Y;7_1~aei{6^Q6%g1Zx
zls59}0jqN?@)h)|BjohH0U&q8?Xvvo^?%-*>+*JCHFzWVHl-zN|0BN0(8K2FZDvZz
z7c+$No{G%{JEJggZD)HYHy+;^ZLyoNBGz%_`P#hO!O^!t=c3oR#yjc?F0<7+ZoCVk
zTl1^TQN$T+WB55WyD5k$h{#9{lPZ4dxfSK14&w7-q%Ryqa?CpZ4KU^Qa42C1V~rLV
zA>^s-qY23z9yW3CS~!Df^Hw7XGGW0Ssk<Ml97($I_VEzr?d5&nNVMizk!<oEr9w|6
zq>c5*iT=2Cpc`_8IcR(J7baMau`7K;LX5Qz-1?q-EuP0^5m7qX%2%TU)~PkM7I0_w
z?~?$G$4V0M?^%$Ul?gV{=(WVxOmSrS<u6<pY)&7M{-Ce3qv}AOQ7rlRls~~0pv=Cc
z7qXq>EO9+YP1McFG0*b!1RbcE$^)~2rSB?@4jzGq6hiawJcSu2lB>6D7o9`+Ug+Wo
zMDJ(kQ=sP;2u5Iv+>}^x?`{I?LB9x2I0I7Z@pI9VpK*Tmu@7<)N{%lw+_?R-8QXio
z$m3o#%vi`_u!}igx_wK=?MPo#$FLHYdwSk#RE<BjvXMK?92_z%+gBza*c9t@DWyD#
zk`LOPt+=HtYvBQPejYO@HCUja`p__Pi9tIim25n3I0C>j$rPiJA$#rf?b~LSKO#p|
zyOxJ6GkR-pYt5p4e=<RT(ueP)1taTF;93R!3N^cv)*1@G?H>=<n`vNh3aoC%P&uW+
z|2W?W=E(Ot?2i0TbmW`BRX3Y?nNjoMFE#OCJtIVViyX_?JLj+G7V6bKn)>Bce<~)C
zRAWDxF#y2U*Yg?Gvg0tbReT^=*iUA4!T#jd`R)+7?$zJAUMB5@8tBB_wDNk-L)$zK
z>VxfX_r+?pw)4*yJM%}1YTTWxKd=Tu2XYI3)w21>)u=m(&i&-NR_qn>;SnO!JDA-W
z+)nK}C(i2G`}>o?1vIR$+a%D1M7pvA>A{1u=>Ryp#r9Q?=bI4ZFJ4I7xXf)SMc&f8
zs2PtNpz`>j)^2N#)Hivr>v@^f_{SP}p-yj8G1U5p8I<6x;4gOAI6fSta`iLsS&1jH
ztRj+FBetJ<{l)KN0Z^zzo=9fq9lu73KQ==i<)2E_i^CCQi3RPVlTdoV6Tu%5^~eKn
zA%Mzgb>XtY3FZdudW#!<k+PjZ=7t%GH{{Fvjs1-ZqOo7cwP!2bxf(|1G2`13Z_rE)
z+@ZFkN|#1q_{#$@9MqrWr$e>JIxGL8p>;_{g7oV-xON&<ykl#|2%U~rK&NxhEI7X?
zgGQlL6&<B~hy-=&Al|`^9zjn7JfcfZDh%GKWQ@U2$(>2-jds^s0h;jx)=#+)&Jl|k
zj>j`)#FeWb@lq=vJX+M50BHUM0t>YL24Seq$0jb$CC10g$12Ig%_S+tC(g#s&BiCr
z#U;hb!6v~WOe*;QB1Q--idN1xp0=dyT->Dpd$u8(=VOmBLyNxq8d%x0aI|3c{yDn_
zYDx&&fry3V1;a(-4b#1baAw6NO@*VBr4@%SP!4|5^wRSAZEZa;#F#Z4@RL?71cElg
zJ$o!jABn0FD!viomnuRoS~xjvIHBVRv8E{1XfZLl-sMjhO);*eL4r$BrX?{5Zc{{W
zGi-pM2tT!1^6))}6#C*foStncpKT<xZMdIpG|FiN^J$cxtOV+#ILYHk2&31H53JQ{
zhp|L6h@ru3eLQ<L2204-s_Zaw^+&q!6G7UwyYaz-&x(Wip5P5vRWfj&qwf20pQG*D
u|CkdR#2X^P8)m_)=KJ{pyv*f(g0rbpc!3%}AV?56SUC}>sH9Y+5&j1ieWwQi

delta 11719
zcmai)Q*b2=u&ra;wr$(?j-81nwzFd!6Wg|}iEZ1qC+Dww?tQ**-Bn%v*0sJ~-KX#H
z>+kRh6ku#TX{d;xG{9JGo#ag}H2>E+<9A98a~M?Y_-;z2zwRqpvazWf?N@$#GIKDT
z7!?U~V;e?&+g2)M=IgQ*av+%?AN{M)2~zQ@cs>4!{}ewL4@SN;i&1AY_H-qq-6)!k
z%~d5bRum%1bgK*F@3Y)(kl4o$TYWMq*;ZNezCG?QQ~mZwHGoU|37xNXiIDMfnigc@
zNU+M3^yavfmA{f~Oz&Ne>DD?E?&B2!#6FvcG9!Q{vWLE^V}&SBGkoz7kJ1agO@W28
zm!Z#Jf9~GC4~*)|a)-1<`BBj!P*p*Lei;H7GsYfiP_t77tUj970XdtQ7E}{!11oZW
zuQNAxxn&^c{{X)cRsUICu`ejB>YWNJX>C2bU6w_?wYBn3NuX+RTz5W~UVr_uB=0(&
zh|Z81kJ3+S7e4so!x5;_FqW8>(jem%NN{E=%3lMiV<=ntz#!KcbI(@(HEFrXXy+gW
z?K8kDo3YfNfvtc=S)*%<!1|dYxBMIjb<&L;!&S;okqH!5#6<wNlRT=)xc5vyoI-0^
z=Wy3qHqkv@V^EbF`N<f-4YwPJ<J|+JhHjO<YV+B!RNp_t{TT$?e5$RqY3h^^=T%Fx
zp;8Bx1CxbN@t5=RZ?G`MnK6bW%NcqIGVh2>S*YHdH-(=h2E~XJ6a1J{dyIi2vDgTM
zkv{_&cLUlfhV&Zn7o+@|6Z7Z&0CU%mu8ErW&EF8W`<#vlj(QdJWhF%HcjBpELr2c!
zcC1zDsq!Y7YNnEOH;A^$yM}vBk;Gbk8K)O5T3zW`7w%Uz*B|w;7jJs-3?KC|+(2(>
z6%k^db$ai*2yA7umzAh%Df?;(AL2*DoQJBy9|FqUB`y_Ck%jguumMGeQK<U+L9oR%
zw%Z_*1vl&ic0eqg&rBW92Kk?0+a~y3L55pou>GEz#tE984H)h0ayY*sWJpLpvG9_U
z&cz@(EPfDT*OKoBl>Oz?sY!D!b%RCix{&?fKIiQ=jFSwGzi%qMDR4zI(7FY1u#7>v
zCV^u08jzDg%yiB*3_{umi6fgA=SR+XV1ew44zq^opj6`FF&VKTW8LA#@)2HFmDhj7
z_~$Mvy|#i~O-hwv<{Fr)*FS<UY+F9Srbzs9F%zfM|NMy>e!XH3OyoiS)4=(bK<aEq
z%*T>?4myJe#=R$l61Qijg6$0>!3_Jfn*^kKG<xOVdOQ>Ow=cfP2hIbcEB~V`=vP0T
z;ngF|4(+WHSe1yx$4DO0A2aC)A+5i2lZaT1PE1L{`Kg@|M-k=_;(O-$Np%$B#70CQ
zLnHIq?=pD&PmT#EnkrvPhF#F0nF63yO5MAX=ZNQv0J9?(+P3g@4GVr}Qm^UMx&l;G
znd?M+c1!~;N(-+hsleP%xOTVEY<14H#9C}EW%aa;vm%0?6WSyKi}@zrM}(19m>+1@
z=?)qk#&4Avx-ed|WwGT7!8pqigY&m<<Bz5tjfaBCBUCVyW5z_+b|`of>Vz7lJ=EF-
z0>yiL@@{;v!`cjUXWTHhCr|_m%z&#)acE|+s$12H;hk2n@A&KxD^%9q_B~#IFbKoQ
zAOGG2Bedvr`dJQ!3|vkQ)N@#6iS|6G@ssASBo!0pbgYub6TigqK^w`0`mYI>O{u><
ziH0yz{v+@BmqbA<lsUWXu9cj^TfVFcy{2fUEX0LSzu64VU5_F~g0}<Y%mq4@M|kp7
zw|+hhKfe#p@1UT7@(UPC=QW5s_2a^LLI7gR1h2T=i%IVWVm=pjcat*3kT=5D7r$_V
zB4`RiCI@+A2ol)_mO%G^B69}k#~~_>d1E9JSyQ3VyA6`#!|Nr2)d{7%h-gj&^5TB8
z34f(rJh5bM;IdGs>{s2(!2+9jP%HN*&q%zwr4+I<p}LTyXI525lZoC(RN-i<F#n8}
zk9KZ|HP&_uMyoEt4u-*idN!d*Q)sk_%Z5tiNxMuGahrsTlApzMJ_J_7UCu}a2MU&4
ztJWW;u~`OHupzUxxn27mp`e}cbX8fqmznl!uA&8od}uuu_^z#Qw*dEy*hrNrrQB2(
z`k%p$<iha*N=<LgGaZ^XaH(P~=yHn2SHX<^o7inu2n0Urw{2Z}-!d0zf6$na;9(H$
zwc3!-Jl9BnnZ>yF?vzj}uTXOC5Q0PY61AQBktl(+%e2Y*ol|1Y!Er|LY5%jKl`Y1d
zTy#_+)diBny5uw}OajjvN2){c+RrzUSjzOe!HH`LrX#v!F!jyfT(j_>tlXj+$@1QC
zkiZirKqx5QvttQE%-f9k+8aQkB&2vv*+xXir3E7zb{P_#cn$RV)%6yAr@gy17*_Y0
zo;O@m*h3<A`$3<_3CXfg|CxZ?#y+utgv<|t%tAAtNPgwc`~wo^2>KV%faO5u#&SQ{
z@htRT+;sVv-M_Go+)1(=sSuqye&b9WJ_44JD0$NP=7_|y^Am0AKnkb!ZU9LBC;$FE
zaYC?qIq%>l$x0XS?`IB8!|47Xi0nFiFPGy(!{g>6@ikRdm2^;{6G(k19|Yr_cuoR4
zP*cFXN~Q3}>4E&B5aTI0R1N47z4Zb-7OKD0ksl?ew?Bm|z+aJok0CcE<&{gg-+TY)
zf24*E;U1yQ#oZ>1ASgJewOPI%HGH;EK$(w=yms$g#<&?6qblx@bKWzJ@>Y&<CrNIU
zR~o8lgTz_&C-(S-_gD5m@b|XW-L;Wv%FE9cTS7l7@_^l08Z$`iM^FAux;n*3$ed4x
zsxdqCj$9Uaoe;>jVQA(Tib4029_^FW_pI~Qa$I=PSoZaYJj~3sV{Gxxnq;m6nHgY1
zku-t7PNT7ZwNO-5xpFY0CLllx`I^)tr#UF1Q{y{7)h27*H8P<4`iaX}VfG=^WzG@y
z{Jz)+pac1YJX6K*zy@hOy>1!X`=XV5UjNhcuL)RP@z-Ny=I_U1moDOy$D#fHmsWE{
z{@ZPT^ZiQmR>UvyI9obyBVr$Yc%S~?f^E|izDfe}vCE&3(u)F((+IbYZ-bpRMy85c
z(Qi8<MAwQc=DTy3D6U35gPP(gqms2vD#zfuNi@3~CzyN&ZOufyEP#d-5CqvO6fa`}
zLxJC2A#2>Blg!cBKhQ)Ry|&3gu+-q3ye$7qO*J}t4u^j+|D&coy^0#n{jD)lEHA6T
z?R-h~;>ci(wQUf6Ck9mcL^#=+8lk7Q(`;UL4B16@m8TKBnd_}y7kuh@A-Td^d$ga=
z)vhduokxxdo!lsn&aQYKc?bqd)cE!k26@Glix}97qAs&1QvW_yGW)J+%}kDe(#y6I
zK&(9C^T(~t+P-aDz~Eu{Z=N~Zmi_y=#kWhTs`fKF{=E)(#n#s+J<adAfs>0zukww*
z#1&P{?W+r`ek69p(T5c;O+y37M3oCGY13t0ebTU#(`qOs87vf|)=ijH_R6Q7{rCX1
zpRad>g>+Az#Wu6iM`w9;Y<-9Ck6J4$Ac!gXh{F!N5TKZvi5CSE{$M#$X-Y$QUt4ZO
zqtYTgDtAnf1iiKPyvj-r?+NqlQ0%bsJUqz#?k#-M(F(r|y;lR~ORiw_(_|JVHUj={
z6T=4Fj5CrzN_}|KdoHqG4#>I7Iv6AE;Q@#3tFiv8)3F%c9Jk{&Ph#^2>wF|TkSazQ
z2@wQ`s15PDFRyxFAh=0k6Xwr=sM11xNL8rB#*sosHpK@%7`*b~^Nj7CDjyjT@9YvZ
z4KCbNn7#Qa8U%FBt$%F3j=<KzdM#|3j0BU)$A`{%5De53CY|iEf3Iq&epKt!f%Fy@
z0(DrFp$CJ0GW|bY=fm*3O^8<mZZFpP2`t-bar%4+G<!aR`$2&9up@k2M*7QNj#$ZU
zUeRqXHe2#VD~~ve0AHfm8cw^Kx?Vgbc!NEMsLmP@*%XQ*4ic3(7c;nHJQiydT%K~1
zq-OSKv`)h-!o^#L^pKxq(e_AFNWh=?JqIK#ZQutam2D%rb)8*I-BM72nlOfr;9jHG
zl|ELCI&|GDSqEq0lQ6)$H{Xl<sAIWl!V|N>rghxn1QtxUPZ|W^mpyKX6NSmL!F^ik
zT8p9%g8*}{0lZ83IfZ4$y=X9ol}aGkaIH57j}+pEJI{=)aPvajslBDO8y=`N=6)DG
z(_WOSAIG_4aI)~iYlRK))xb%BPvgMXfaBApq3s(_N@bsJHa#&(n;PAE3`DZ6ZPQ<i
zXKqMB2LUoRAcXx6iSo=RKg3aybj!81+FksavtN_EHu{4tK*+diS4t&3+B*P>wCp2Y
zrlm+H;2pMjR4>?)iI*!gV~XB~2ZQP9zP&t7FgE*fpbmaEed`9u%;7{w(AXCsWOV@Z
zb*&m#M93PW#fxR4a)9UA5uGYcERXX`kMfZC&diH+eCK&4XXUPjZQOZ(*`U1~0yOQ7
z=KRoQY>}muA)2(qF&!=?MHtPqQ1QEj_{xcr70$kYOef$q;yM&LBHIa-t0C>no?aw3
zWb&u$w1?aa>Dmhj#GzDY1uM7GmD7#R9#5h$`mIC7z=~_#@lY@FCCAzs-*h{*>G=0l
z1oc(qo>ru652xY!1$t|+L2z22hakQ-2yAMo<SxJ8#>Z6m<LkNQVuNb!<Bx_~dnjVN
zg6i_f=?3@+Jji-?lmCPZ!y!Awue<~JIxPjd9Q2kufVcth+BSGWPipD|WpZFnH#yaR
z9S8Sw)6W9W+E_w6&^VV3(ZUF}9-fB=ri2OCNEJpau#>=Ve{NJADKrV{N%W45$nTcO
zuB}SHBjH@TTFd9X**j3D!o|Wlk6e^N($sn^O+)J3(*ghTH^M7q!v?bcKf5#VA#!uM
zw>zqUfsS<`?<K#cMz3b29O41fTl*XM1@2D44unDaJ1N7BAz3j7`yNdbdwd=00_<X>
zuUJH1p+3wv!!*_@5mIXtvy+utdCVdJ$#P&+Bu6NW_C5-RPBrFq{c8m|t0vB9;g48z
zMQH0zuT)ep$a+x9+f!66hOzbuyqR|zq_GZyU@RjrZ*o`7d`5V?!RBJl3EHDs8c%)F
zbGR?h4|97FE|gIZYOV_jO-krqBzV0YK@PJ+L0ltP7JH2ax0I}xe@z&{D;4(44m;aO
zv$>gyv0a|X8!;x44i#sz{`>n{KV^;J%IeR>RckND^*ki^mU_={dVn|YqOa1?z4>1W
zC{8Y5kh^!-(wR1W!>j1?fmVDfEQ{o`+q{oq0gA1`mNT_`S1Y~2cNo1VSXo)lMC0fm
zu?7lPIz;V>Q0J?zQrxo&;#g;RE7|9RTwra0pBw>Pz<Z$0mku?Vy5;qH1}AoiyE^2l
zf0r|<>-k^^;uj|qr^@6Ky3%_mfciA%GcyM$UM+vvfOOI0aGBW{-~fi0n(gn!z@fh>
zd&_XTUF?O}`(1700y^cRV6x*y)TuwDnKKjL4Q8NFs8TjH`-XPR%NE9N^7J)kQ=3H{
zLk(|zrdlaf<dV<DA*l5BRwOx<98xNN$jaAzyXVtPrX4~pHX7gjs<*M0jRPtC8vF_z
zv}@bS*tTmAxwaWeR|wcuDY)d0cOd#fKzBGv<9gJ~0}hX`5IeQg=LTORP{<<|46`H_
zo6(XT)H(y8U8Dc;mJ~62TU(u(NndN9dyttCrLV(G2@F3{2pl|f2x^ANUgZW=Te{TP
z#tXIk<!pPj-d2={(tmW)+5Kn;G+_Xl6?Hsc3*lrHIt4!{m`nod_@v<nE`>vx^Fo2m
z**6bsAfn&=@_0qkh^z4MALopkS$Z(lty%bjEYMD3N8T?=xNM_?u6p&%DO`F;R@l8n
zSkZ!u*g+zqbTmj2{j)unF9trV>j(bz+wgT6Via??QR|3v{J@GZJG<ywuXmshx~jRF
z5PvMrdCh6T8a9BvF&LZ`6jq7(Pf0k)%hQt_sCr1yCI7yfH);BTx8|0N<#z)#QcF0T
za&NDC#?pFJe_Wjvu%G6kVm6Xe54JuPx=3Xqw4Y;{skbJpL8<V$@J0`hH+65%>fUdZ
zFpRI*rD=}L&0w<wp8&(DqYC)JybrVI^GfmNv%+E<Hyf{l7BJ`sVq4A^5R_}kmtqMH
z{W34id{bT;T!HfY8|hSHk$kvrH)agyl7^V&UN>bZ&&Y!Wi3NWb@3Qsv`m)8hq_)XO
zmMKBqIYX?Qg8`b8XVWh}P%s`C#J4?stC`e2r(kG=sDIs{`1rP(`U#};oYhus$DUi3
z%cWL<pk<M?$GHAZHT(nttakj=qO4IdaH7L0->xXou$ufSN>q>$*~{0Jq6Oi})8OR3
zy|MT2!p$m@V4F#me-fB3+`c~<Dx;t?K_#YW2=sK+i%+CFtt-m2|0nrlmsv%=RSdx{
zhbBoZ;R<GZgo|-nI}Kzr#=4_1X?$m5T`%e1Ujj5!q`0pCLWb~9WR&9)Tt>#pGKc!~
zqAPUT?bFf@GRT*nst)?JTZR)vi8&42X&cXM;N=L#12*23-YfjXoXx{;&Ugm3ZtOk6
zK%fblS!=e4rI8Hi)n-vu>gKZm4%x?>tdf7wrP<|F9>v|*4q0eYM%c}hK1I-{kGl^I
zF^*O46}Wl~x`}40#LvdnX%5<GNz$UnTMl7Cer{_6D0h=TFsQ+KS^qCRPV2}!Y)YZ@
zernyxfbLuD#zS>rOc#lwHaDnX$fyM*T15v)@D>w{ru}=FxniPKGn98v0_)NCMX<EA
z?9F($P;8ovv8bD|`ug_re|e%@%pg~C$(d$mlgn<pi?d)b6GJJoD@`TUFSqR--#ImJ
z%ElBc_id}RAU<`-?M<-<J_%m-ai%zql<oFL6Q?VnTZW7vBg`k62U)S0^+J#mh1|4F
z&=qcLIJK>{0$45&y#GE}G*q=AW?-sqe62>eS1YWIw@}0s<-}=U1a$)Mz&#Bd3u6#0
zzH`=WMI&3$AbN)F5>#B9`DLpCIs*}|s>*(*K3g)L_NAXyT|KkFeX$cQyOwI~F1o)5
zlM3hE`bMK|o6+GRGCXP8w7&hL!~uPPzH#J~g%G9g#-1sAH`DkeJ&;jRn91p7zQ6Uh
zBw>?gXBUaInEPBbZ|<))g`&O;N860v9?=AQ%bMHE)q7mKT$7z0F<{qjDFHAv#DUa9
zWz&I4APZet2N(pvDX81+G>R{cBpg$H&6to6JBBJ`jWMbK$s_Z$AhyFe81Mv-5PYWz
z3>iEss#1kUc{EsB-I!t_6-mob_i+KV;Y?LLOtI5hHD38h6}F+`7+vZ}I--nn^z8uh
zwG|NO5t5a^C$mGy^1mR{jAPrguz#Dfkt=|<ifH~V|H|zL(q1XBx(dB9+80Hd`hr=G
znv<cA-p_})EKb%B>xQ;Pnup%w^3Fs>k||vgrOaB>5s`@T!th;6sVa?X<<KRwrItav
z$-r{WEGEJuITj9Vx`rfT5DjzdaOlXQfSP1!7i$luf$pVe{!ZbrWInOXn+<bT5XJw+
zk^(ko**J0ny!w{TWZb6|$%ctoj>X>og~tI8G04X+{7LS54|@dx`H6X<7J}yv=E(;V
zPYj*=!P<X_43_Idb9eQF?MY0tDqN0Lf?6^M_^UWspXCYx6*b^Z6q-x9AvF+6xzwTZ
zq2qcsUHRK=w=+X@HG0Yva4?;=Kb~{K#*(54;1ah13Xo94&g$HplllEZ%)1&!D+w<q
z>fG|o0DB7pnZgJ{qwix^*yH5tU*aa~r?f;UTuPR5=br8eF6a3}3u(=?)xe=O%wDRR
zr0gYLC(C!%^gFZ1EJ4pS(3~}*mh(q6PMH9K);PgAHgnJjHdzFY#Ly86$t9i<@+KdH
zCe&{QNb^iWR9iaP@ryTJOt>Vn`dAP-^$*Tp_7PAmDQL0}`EABp{@EVeJmi^E@uOHS
zG}my4E;<1)f3Kb$H=ZV%+*wP_lWRLqKy8;@%5E4nY01ay5MFeRWqEp{w<<VjXqG`j
zMHY$ky#1jW@*22MO*%<ZQ+>Sg5lzYDF`B0raLhHs!wP}>-4qI_Y2H?~e@xbG)9QwO
zIaK?bJJExyL5krO3D5e+*P!;2aKjycXG@0|9?_cy!zxPD1CAv|6qErs!l~MjxmWXg
zFsr|22GZF%DpbbhPS=1d`VA^=k81|-wpVF8sR+q6dI5@Q+zfX*_}ES=o34NxY}bzj
zv<}LZUcT(CnBb~j)l5UVMY34(0dYscu4qV!MoD{I#_Dkm-bpj|?ii}_NB^rFY|*Uj
zdGOkSo#^qBL4;yLP7wAue(|ZeYuOwiCtj25f6~L!YU~)E(LS^rY+eO&Z0;FwP3U+6
zVyPu5!c>q#>@7<>XiETX;s7c{rt}LF2x5Dh*3Z8LBgnz>9l=$-T%L%Q(UPVEe7{-N
zdD)p1(S&1JY3sAo_`yruRrb4W%@c!a?Yc8I1YT&m!B-Gd?r2=QqpKfs!uB!uY;2sa
zl-lNQ$tuPKVzAIlTh4qYK8$p340GK6E$SPa^{CgWw^gkj3f}GC)D9>lJk??ZZUw{-
zmJT`1v3%r92n!mn7_kg}I&tv|vcbs|jFEFDHoV$RY)=dc?tZ}v6R0ywy64|U%;R8s
z5;&8D7v<hm$S0wRRQzRXOxq}OX^fy!Xj*bjOJ7q+Ujs4ztukTV_@f8Mj$a>VCM~g+
z_w}o6Ezutam%KTzvZ&U}RMhGbxJSFzeDS$Ru9Zz!Bt~eTOz8<)kFq;kp&5Syg+4?A
zgF<zewz+6Unv_W*W@cI3eti}pN<`j<#!~1A-e03l;ADUCFX94ZWNbHqN~YMoOhLd6
zCMaCb+ZhH;o6J$!vEM1Fydcyu5jHwJ@CrvZq?2SsHFNEfEy@ZNc5ibSm;$bi@<a;l
zC*EpE+1y{-p+u7iE%(NX2U|m+d<3HQo&Po@Lm{6m(@yL%)+Sq3!C$Su%j!@WK9zOx
z)NQRq<T~cic$OQC^c>K@k%)71c5Y1PR={>|jf^q3%WE-h!I4k*OKdIgZsQtyX<v5T
zZFSD+USxe2dN4$5$QXVVn5zfnX71_iMEEZnh3KOfPB~2m5SLr2ca7Xa{kIfdGOp~)
zHQ-e!K}KsS(9nI$0WAQr$Myl1>~%01dkyg1BJ3tDO(_jV%KB2f8!ilLw9B=rxXm8j
zDU0a?y3Qu@Xfvqv#T-ZQ`it$DOh34~-wc*AA9BQ$mP$;8;|DShyn*8^xEMdY7$3B)
z7$0n93W}ARjfI0FE#M4{@c+d{XJ7-MT>r=INsB!PBLF&4txg1!Epl-;mM5N3>=^YK
zag#(QG3xa$<<<<7z;NJi6MsO-B}oF##3lzZ7QzOxHZlh>S3bnT25q#4$>j9LDkC_g
zdMfECI#cKQ*$(jZ1NACQ3mgF}SaR5E;gu@GfKRb~RwEOlBh^r`VDYD<6#%B^@8GIf
ze!TbzIN$@1K4NZ>O^72NLk$zc8DcIwsJa!x88{YeKKu-}lnHFa?*fhE<v2jQ7}8mB
z%~GgayOfa#`Zb-Nh}rePtQ13_qyZourdMenjx1anVaWxGLhhC)nrYgMM#8{Y)`3o^
zUHc70aDYxm9kh)Mr~1?mRUag-7Y0>9AQ0D)92jj?*^cgt3cihAjTQj?H+6jeKJ#oo
z67@45)(n7&M82Rcqe!UyI~d276-)?(;|7HOKQEH;AH6ih8O{c?xK9XbebNLy3X0QR
z04jMW*jXkNCB`U`Oby~*QC!VYoa-467}Fe){06IQA@tFnS%2ePb5!9sG_e_nwLGB}
zHjuQyX%9KJeh8G%k!#>02PO=P20M``zP@(XpMF@BG`Hn;Lrj3S*9zxyQ|yX+d-Z_$
zr~fI=(?h5N{{D2YeFxR&y1i<2yBbEKjeKXN>BW|>+LA??$P+-!6+C)fcOVZyiG#y0
z?7i8$r_Cj|&am|mupc;v_YGQU+TydP1U51C-c&QpB40&}UO(06vfW-(c`_MB2yTJo
z)KN^<)Ax$ZHxBN!%cUamfJZKXfKw}4gU=IY`(5^{ciaBm9l3GnyN5INSmV6bCWz?U
z>CNwhrZ(Vy7bf-p_ob7|X~%0nQq9kAdr?uYJZhy3z3n*{S@f8(_T!pg(o%>^3|#3W
zT0T5{DCQ>=3;}Cf6tw{i-o_x9hY+Vvq+(f^YcTZ0a;uxPR*JvQdh5^fjqSRsn7s7h
zMi*Ixd=fvqzRbD{e8w^6wiYGw*mov3Ie2J{q*uD$IRVZWhAd6uOYRALSZ*A;37eAL
z)mj%jabssDO?>F67GLgc*CHpZf!F|mPKg);^5$kVLtyW#V-?LLhz&(zaKRmHbkc(`
z*~0?5u>XgCy#x|#ZKL^d)cv~ElN)-8g`9|<oaZN2G1Z_+{tZ^~L=y73oeXO2gk0@-
zC-Nd4&B2wraSHEJF-qnq0iAl@V17nCxnajv*)OeQp}`6{-};a@mklIj;BQ;f!R6#{
zvF<9M+hyP{#<&XjK!uJ?Nw(A@P1#@Ow^$?9%!FzgfWYmrEom|c?f&06Jzg+8w&6|A
zEJO&K_xk}gcSuM{q3l$R$24ego7T1r!R&3(%KcBpg8V|~n)<(ZaN3d|`N!lM)C;Lp
zcT#dq|2xV-Koi=N(idkb0F`De5$~o@8?8sG2JVV+SD7ydT(@H4VeMIYKcPtD%ARed
zs%he0ZixM|YQi-=lOFVZzUi#peU9a0gG}*t2>}&P?!Kz`sYt1_n$0r(8<*EsCG?6*
zS%t%L5NFbnh8$<oXc*SsO5%{q0!b=07pH!b=77?DKDu}pJNIxIpy`x11?$BhPSu6c
ztr`?DPNUzI%ry*xaAzII>Y_*0T{&qpRO;y-g67nMxe*R_p9T`dbcDVlk>^!irv|}}
z2Toga8!nhBM?c78@Z_AEmIf}?OtCavZLFu5>em_jR87}eUke&q(;c^P19Jlh-Y)-4
zYv_iw^oy7QiMrXcK=};4=mfMAySb9y;^&0spYX?2+Uc(YQ~_W6Pi5ZNhnFNS5C;$0
zXftHkL}t+mwHZA=h=W<)aL!tL@jcE~D|Ww=9_?T)1-rT8@DyLOZZZ(Y=*+NU#k8s$
zI)4%{+AJN4+AQ7ddl5PO6Jwi^Yu%bBU3`Jx_spQYpr_Hyz_A!g!V2)=g5IKJwrlec
zMG#zOQbxr^A5h4OkD8S3*mqMjbX@CKRZ(#WoqcP3#8{l5spOo3P(_}lje7fFOBlo}
zUCF}@*#mu9#nO=MT9`n(41996yBsnQL425PX$G7HGgjm(!)Q#lN{MW5Ke}o61YL{W
zJhR^|+P^o2fo`I3!fBM30$Kq*6>t>lYLkgrOnV{#o`(%E(wMz`p8c_OTV>^NI1ahA
z(?=ZPtPbj1i_6I`?42sC$xXVf548G{2|r40L5n{gMXRe<z4B2A#&`gEA}{(YI=8|(
z0wLF5{?_8&O|pUh<+>XXeF~FjO`CUS3&A=2Z@)-jfIZm{M}hOIS=qyAnJfIJIb>M>
zx#Z7)QmGFTrS=K7*g!OBLmE`O4e&-|`6eVXG|SN@4m6cbd4EIx4MofGi^Z<K<rA>E
zw&u1HGX@_`tzYcj>w(?6Ns5pG5vtR@HXP=`dpuBy#A;!s*y#{X{iGt6lTp2e<q=JA
zZxlHQKumsHw?F5`1tC=XH$l2bDP#Pi5p($Bt#=~YuYxQs@)`(w+&KvZeRu^EfxMMb
zi5B1p6tq3p@}K4E*`c!3d^lzZ1pi2E0I0|>wDPBnYvDzHBmYdY$<<%jXMHldm33Ru
z##e%S!ehtbCe&lNrl=s~P2vmWzw;v+f)b5UfngLa8KvA+>ojV@(>o=KzG_*T=WVTI
z+G@MflNJS5mpRmUao|VRYuV*4A59r*kM25V;}jU@X9UwyBAD!oCsPE@!ZO}b#*<Y?
zZXKzB6ngMDyBcJvn=+ce9P}!zhuP^qel?!)J{-7L95-~$xW&zG3X5gR98kSt_&C^`
zz_mO@xDuf@78qbG_s7oEvYl}C_Fe;Hy0gpsnQu?aTeH$iT_O`DKnOV2P{CI8zy<AT
zk1okP@`~m+V>d29E{Q_mu+~SYMT=qMl9>?LP7-OHlS0z&78$SplqfDsT(5z-X_HD<
zGdC%}^#b2$o-k73kumbm2rep`|HH@-sKS6bhP-jn*O4E<S}sDWgb!z0ktfnApY>L}
zvWV(jGQyhL$`I=r$3V-EU)kyx@cocIMBrMRw2YopI$PB~kzYoeOh41|dxlzY)o9+C
ztluTWr|=L<8dC`CC5(Xmb|~S~nmxk;lH$Xdm6!Up+<0N($%4I`!{V$y;buq-SWRi3
z$(?!thIFiZME=*5iZ<b6YoQ}7o?<=;xp>9<LjY?;A|IlcK)wg1f?*!!PUhyn;bv5Q
zI^~d#SSstV`*~R0qVoS<ao$grZl^$GbKIe<gHK3jbA8}eJjN!@B1@GT$6jwY8#$)V
z;y2gOkD#c3R0mMI2&>FD@&pG1gMvkMwI?vLlQFiuh4dKRW_mD))j@NlPoO_t52pQ&
zJ39CGu1+T{1k?uf-+l95$b{SSOS}$c9n3n8b981XQ_91IuX_Fz=UT{CSZ*6^P9<)3
z%fok#rDiMtw?El<TMsn*DwG~%s!&wG{w_Q*O*MRkrV|{?$x%*CFk_<v=1+y@xCB-?
z%$R5P;Dc^JkKq4GkHh+dX3glPei>vWW0z)C?8olJYx!1Qw?G?JmsGQ5#bmXny#yA@
zfycQC5i21?cJBhe!J(Xoo*+DrZXrA30!w-WyzY@GPP0<JSa;k|=ycCCtn4AmHa!_h
zp3U<Ye%no}+i5KeZm1>#3ffDDli(o6S%%7{#J?Y3l7D0lkV_KBCj(?ERI8%5*Aol}
z$zX0GdIpZXwh8i%^mG$Wh{j>0m^%l`ac|<{rQKAoBZ6y%{FHam)=-Hb#br?UoY>MA
zs#>8<wFFQ<50LNGy3rg)0zIjPA_L#3!VXlH4$9vHEhyqPnksmJSkwnczRqg#lH|S7
zi8#@k&4;=Q!x~e=wjAj>5lXJqi=9})Bkx1StN!3R`X3NB$2pf4?YZCg=l7qh@a;<g
zN3WphUP=?ep<^FM$R;rCZF_JnOzy#F4FPBY@sEIGR}m0aI@q{HPLUH8S7E(-@PMxi
z@Pn!>kcYR$(|BT_KQu`l)b{wIk%3^TvZb-K!{uSqys<Yg*F*v(?-JKvHBC)<D}`n0
z(u@5d@DnI~ePS%e&w7P3@gP;dB|-bR#p#r>GkbTa;<Mvd6ZWhasME>7+JDP6B8B!~
z<5_LQhSxol2^3y+L*n#%i927;1*;y_obvNpIuqm%WihhA7S!ajyYkg5_1h8Yhl)=8
zUtOfpq&685vVwLS;%sRrIOsWVEf?H7TM)@*&<?voIzti$!I9Z!g<FCqrrevbCiR86
zU6I~h#6sO$pJtMRz!rCxV~XXd*TQTq{8XQC1Hw8hkD~mxpRe_*(L8V71G(v#+OdxD
zHcH^RJ>?CsrZlqNAhUh2&tt@_Aza}eqvHS<$n?T*8N+=ro6Y$TjoA1&g&*o^B*)@~
zQ1)okmF3IO0|V#8FAg3<B<uX{>d8vuXF^EyJZu%~dDjxWn7~Qm2Y=uvF^s7=Pu+<R
zM`_cE`HJa4g&Ahr&z6lZE3$87*OGBu8a=z^g4h_?p4ruD;hH}9q64>ep#LTRRh8TN
zy^Y8nwa2{E?fr5;`+j3+)=Fiyp&wct)(MkkbJN}7|9C>x4J)Gd9+b%)HK8nYb9-|;
zy^OG0Zf&5sDA8)-<H5O&$<@7D<z`q1;2&~_kk{oNDbfcupaH725_5ytnfQ!Pt@R@H
z#WDdaB9kP|eRRX$r2_vw5A1BF5udV*+8A@W*c7sXu!Qpu5pb9FPzU@O95k|VTNnpx
zaaSe&)@DW@ihUS34OYJO@Ng00;Sv5`iL+Q;k!bK7rbLY+po#Rx3jcEKN-^dNzSs5T
z$B+Fz+^p&cC-on3a~_uRW0y_i>|3$fDP00FLF`l5UJQM=e(It6-Eil{fFFbPjZsLZ
zuLVCRbA!4s&(Yj)nO%81f<Ir*fIb^jxRMbWBjBGvLza=xIR_Vo0tT+i@arSrPPK`@
z@<Aty_w|Cde@C@jc=;+5bI&fhjZ#h}R_|CZ%Lj1XP{rT~KhG}~j#|fBvv-`mG78g>
z-oVH}7t&bYz_5X%Pd(s^5VE|X!G^8hotW+m22KwmL56}h{T)o%Qmy*ZPRF{&b?r*b
zw&_Ka5k;<O$|iOYBVbSv?Dx4VUt=_L734O=vMzALrec;ZtR=6+g+<&S#oz91fbkBv
z3ax4sO7S?}P*~bthG3;Qu{)o8?=C>fgd|qsRtc0$)f3Rp#+p_2<+6|Lq6N)S6-3M-
z?RRY*Hu%s1JOC1T@RJDHf)i+9eW;h3>M7PYzUNST&Aq52U8_d}QJog-OOEO3Z-vyA
zjQgD(EPHA^r_8{FsHG*_a(+BgsbN?vFUgGOtp<><_!5+H8#y-x)H{l@9BV<S%MVNm
z!nAgn>I0J}YbQ&lOYJmXNmftLY!sP*(^4--jsIm>bwMi(igoZb2Fm>?SP$!NTUU%n
zT{k^bnU2`*fJ(%(^>(V9Nba<oPHqaDg?F1`JvgoHV8_+&UG3y+F>QEycj`3I;SV2_
zn+;Q+X-1|9t>@xL`63tZ?X#=t(5_k(!&JrzVAZe7{gWHf=F8t}obyb4RX4Ylh6kGe
z_3;>otIFN`Qy%Iz@H7p%`PKPLu@>GlV=j5}c`$rkjNzm8E2ftT2t^w3Ogz1B`){(~
zYcb#{_Cm5m3>q^*z;gkKn9voR0Q`))QQ~b6QVF@lSEet~V26Hdbf?E(MmWe|CoB93
zNWOO5Gt#3-AC)$3y;#G@Rn|#@65kd6gsywS75JM%@oHNPXSpw&h3cE)Y@qtYq521z
zzFsKov{@hfcdv2YRggwBukJ)m)7k8c1D5TA|1gA#f~lkjra+og;CBGiNAzbOx5$!R
zyhjo#ga1=Par}C{^$oyBJG#%}IjhMYBV-ZH_GG%4sLbIVJG%Pap#bsERNBf1n4_8m
zizK@w7pJJW6pIuKmn4@2rz8)jC>JL$H@7&aqy&!;iNOE2!~l+2!Q9@`)ry3Tlau6s
zA8rWeW!b=uQNs5B`J2Irq6`zTM$?{<TA0Exp<=5j?LpWNAQA-Rq?XXZi__x6%j0+V
zcwN+7RNhEtWqd<58N=n#7D0f@<zrNuV?A2)YV^}p_wya)qq&oSnX5qB)I&q7fVZFo
zTSQEdVfz;$T@-+Fks>=;g;1@98LWjPS^)#yWyTI1dg~CA>LF|DVF>F%a_SKb>LIr0
zz<Uycgwn$~o`ygW22l_uleqmoB;37@Iq7_OU-6RtFuU2*o_1{X<uH1vid}bX)H-Lp
z6=ipE>#Rc$YP2->Zw73n_V2fCH22@9Y><%lVc;AQs2wZs?mvER@~A(-bL7Zv;nMoP
Sz=+`3cv;~nDJ7Mq;Qj}`I)N$x

diff --git a/doc/Projects/2018/hw1/pdf/hw1.tex b/doc/Projects/2018/hw1/pdf/hw1.tex
index 4e6872d7d..9b7b511d6 100644
--- a/doc/Projects/2018/hw1/pdf/hw1.tex
+++ b/doc/Projects/2018/hw1/pdf/hw1.tex
@@ -134,7 +134,7 @@ Aug 27, 2018
 
 \subsection*{Exercise 1}
 
-The first exercise here is of a mere technical art. We want you have installed 
+The first exercise here is of a mere technical art. We want you to have 
 \begin{itemize}
 \item git as a version control software and to establish a user account on a provider like GitHub. Other providers like GitLab etc are equally fine. You can also use the University of Oslo \href{{https://www.uio.no/tjenester/it/maskin/filer/versjonskontroll/github.html}}{GitHub facilities}. 
 
@@ -148,10 +148,7 @@ IPython/Jupyter notebooks invaluable in your work.  You can run \textbf{R}
 codes in the Jupyter/IPython notebooks, with the immediate benefit of
 visualizing your data. You can also use compiled languages like C++,
 Rust, Fortran etc if you prefer. The focus in these lectures will be
-on Python, but we will provide many code examples for those of you who
-prefer R or compiled languages. You can integrate C++ codes and R in for example
-a Jupyter notebook. 
-
+on Python.
 
 If you have Python installed (we recommend Python3) and you feel
 pretty familiar with installing different packages, we recommend that
@@ -216,7 +213,7 @@ We recommend using \textbf{Anaconda}.
 
 \subsection*{Exercise 2}
 
-We will generate our own dataset for function $y(x)$ where $x \in [0,2]$ and defined by random numbers computed with the uniform distribution. The function $y$ is a quadratic polynomial in $x$ with added stochastic noise according to the normal distribution $\cal {N}(0,1)$.
+We will generate our own dataset for a function $y(x)$ where $x \in [0,1]$ and defined by random numbers computed with the uniform distribution. The function $y$ is a quadratic polynomial in $x$ with added stochastic noise according to the normal distribution $\cal {N}(0,1)$.
 The following simple Python instructions define our $x$ and $y$ values (with 100 data points).
 \begin{print}
 x = np.random.rand(100,1)
@@ -224,11 +221,11 @@ y = 5*x*x+0.1*np.random.randn(100,1)
 \end{print}
 
 \begin{enumerate}
-\item Write your own code (following the examples under the \href{{https://compphysics.github.io/MachineLearning/doc/pub/Regression/html/Regression-bs.html}}{regression slides} for computing the parametrization of the data set fitting a second-order polynomial. 
+\item Write your own code (following the examples under the \href{{https://compphysics.github.io/MachineLearning/doc/pub/Regression/html/Regression-bs.html}}{regression slides}) for computing the parametrization of the data set fitting a second-order polynomial. 
 
 \item Use thereafter \textbf{scikit-learn} (see again the examples in the regression slides) and compare with your own code.   
 
-\item Using scikit-learn, compute also the mean square error, a risk metric corresponding to the expected value of the squared (quadratic) error or loss defined as
+\item Using scikit-learn, compute also the mean square error, a risk metric corresponding to the expected value of the squared (quadratic) error defined as
 \end{enumerate}
 
 \noindent
@@ -244,7 +241,7 @@ where we have defined the mean value  of $\hat{y}$ as
 \[
 \bar{y} =  \frac{1}{n} \sum_{i=0}^{n - 1} y_i.
 \]
-
+You can use the functionality included in scikit-learn. If you feel for it, you can use your own program and define functions which compute the above two functions. 
 Discuss the meaning of these results. Try also to vary the coefficient in front of the added stochastic noise term and discuss the quality of the fits.
 
 
diff --git a/doc/src/Projects/2018/Exercises/hw1.do.txt b/doc/src/Projects/2018/Exercises/hw1.do.txt
index c19d24d2e..7e9b014c6 100644
--- a/doc/src/Projects/2018/Exercises/hw1.do.txt
+++ b/doc/src/Projects/2018/Exercises/hw1.do.txt
@@ -5,7 +5,7 @@ DATE:Today
 
 ===== Exercise 1  =====
 
-The first exercise here is of a mere technical art. We want you have installed 
+The first exercise here is of a mere technical art. We want you to have 
 * git as a version control software and to establish a user account on a provider like GitHub. Other providers like GitLab etc are equally fine. You can also use the University of Oslo "GitHub facilities":"https://www.uio.no/tjenester/it/maskin/filer/versjonskontroll/github.html". 
 * Install various Python packages
 
@@ -15,10 +15,7 @@ IPython/Jupyter notebooks invaluable in your work.  You can run _R_
 codes in the Jupyter/IPython notebooks, with the immediate benefit of
 visualizing your data. You can also use compiled languages like C++,
 Rust, Fortran etc if you prefer. The focus in these lectures will be
-on Python, but we will provide many code examples for those of you who
-prefer R or compiled languages. You can integrate C++ codes and R in for example
-a Jupyter notebook. 
-
+on Python.
 
 If you have Python installed (we recommend Python3) and you feel
 pretty familiar with installing different packages, we recommend that
@@ -68,16 +65,16 @@ We recommend using _Anaconda_.
 
 ===== Exercise 2  =====
 
-We will generate our own dataset for function $y(x)$ where $x \in [0,2]$ and defined by random numbers computed with the uniform distribution. The function $y$ is a quadratic polynomial in $x$ with added stochastic noise according to the normal distribution $\cal {N}(0,1)$.
+We will generate our own dataset for a function $y(x)$ where $x \in [0,1]$ and defined by random numbers computed with the uniform distribution. The function $y$ is a quadratic polynomial in $x$ with added stochastic noise according to the normal distribution $\cal {N}(0,1)$.
 The following simple Python instructions define our $x$ and $y$ values (with 100 data points).
 !bc pycod
 x = np.random.rand(100,1)
 y = 5*x*x+0.1*np.random.randn(100,1)
 !ec
 
-o Write your own code (following the examples under the "regression slides":"https://compphysics.github.io/MachineLearning/doc/pub/Regression/html/Regression-bs.html" for computing the parametrization of the data set fitting a second-order polynomial. 
+o Write your own code (following the examples under the "regression slides":"https://compphysics.github.io/MachineLearning/doc/pub/Regression/html/Regression-bs.html") for computing the parametrization of the data set fitting a second-order polynomial. 
 o Use thereafter _scikit-learn_ (see again the examples in the regression slides) and compare with your own code.   
-o Using scikit-learn, compute also the mean square error, a risk metric corresponding to the expected value of the squared (quadratic) error or loss defined as
+o Using scikit-learn, compute also the mean square error, a risk metric corresponding to the expected value of the squared (quadratic) error defined as
 !bt 
 \[ MSE(\hat{y},\hat{\tilde{y}}) = \frac{1}{n}
 \sum_{i=0}^{n-1}(y_i-\tilde{y}_i)^2, 
@@ -96,7 +93,7 @@ where we have defined the mean value  of $\hat{y}$ as
 \bar{y} =  \frac{1}{n} \sum_{i=0}^{n - 1} y_i.
 \]
 !et
-
+You can use the functionality included in scikit-learn. If you feel for it, you can use your own program and define functions which compute the above two functions. 
 Discuss the meaning of these results. Try also to vary the coefficient in front of the added stochastic noise term and discuss the quality of the fits.