presentation.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Launch MLOT and other methods"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Basic run to check if everything is working"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Namespace(SA=False, adaptationAlgoUsed='MLOT', cross_val=False, cuda=False, d=70, detach=False, featuresToUse='surf', lr=0.001, margin=10, max_inner_iter_grad=1, max_iter=2, new_space=True, numberIteration=1, pickle_name='test', reg_cl=0.2, reg_e=0.2, reg_l=1, reg_pca=1, rule='median', save_pickle=False, specific_comparaison='None', svd_solver='auto', test_exp=True, time_cross_val=1, type_classifier='1NN', verbose=False, which_dataset=-1)\n",
      "A->C .\n",
      "     40.3  0.0  MLOT     2.62s\n",
      "\n",
      "Mean results and total time\n",
      "     40.3  0.0  MLOT     2.62s\n"
     ]
    }
   ],
   "source": [
    "!python main.py -n -t --numberIteration 1 --max_iter 2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Test all the methods"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Namespace(SA=False, adaptationAlgoUsed='MLOT,OTDA,OTDA_pca,OT,CORAL,TCA,SA,NA,JDOTSVM,LMNN', cross_val=False, cuda=False, d=70, detach=False, featuresToUse='surf', lr=0.001, margin=10, max_inner_iter_grad=1, max_iter=1, new_space=True, numberIteration=10, pickle_name='test', reg_cl=0.2, reg_e=0.2, reg_l=1, reg_pca=1, rule='median', save_pickle=False, specific_comparaison='None', svd_solver='full', test_exp=True, time_cross_val=1, type_classifier='1NN', verbose=False, which_dataset=-1)\n",
      "A->C ..........\n",
      "     40.5  0.0  MLOT     1.62s\n",
      "     40.3  0.0  OTDA     1.19s\n",
      "     40.5  0.0  OTDA_pca   1.58s\n",
      "     32.1  0.0  OT       1.06s\n",
      "     25.4  0.0  CORAL    1.79s\n",
      "     38.8  0.0  TCA      4.19s\n",
      "     40.9  0.0  SA       0.34s\n",
      "     26.0  0.0  NA       1.26s\n",
      "     32.1  0.0  JDOTSVM   5.24s\n",
      "     30.5  0.0  LMNN     1.44s\n",
      "\n",
      "Mean results and total time\n",
      "     40.5  0.0  MLOT     1.62s\n",
      "     40.3  0.0  OTDA     1.19s\n",
      "     40.5  0.0  OTDA_pca   1.58s\n",
      "     32.1  0.0  OT       1.06s\n",
      "     25.4  0.0  CORAL    1.79s\n",
      "     38.8  0.0  TCA      4.19s\n",
      "     40.9  0.0  SA       0.34s\n",
      "     26.0  0.0  NA       1.26s\n",
      "     32.1  0.0  JDOTSVM   5.24s\n",
      "     30.5  0.0  LMNN     1.44s\n"
     ]
    }
   ],
   "source": [
    "!python main.py -n -t --svd_solver full --max_iter 1 --adaptationAlgoUsed MLOT,OTDA,OTDA_pca,OT,CORAL,TCA,SA,NA,JDOTSVM,LMNN"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Run of MLOT with default hyperparameter "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Namespace(SA=False, adaptationAlgoUsed='MLOT', cross_val=False, cuda=False, d=70, detach=False, featuresToUse='surf', lr=0.001, margin=10, max_inner_iter_grad=1, max_iter=10, new_space=True, numberIteration=10, pickle_name='test', reg_cl=0.2, reg_e=0.2, reg_l=1, reg_pca=1, rule='median', save_pickle=False, specific_comparaison='None', svd_solver='full', test_exp=False, time_cross_val=1, type_classifier='1NN', verbose=False, which_dataset=-1)\n",
      "A->C .\n",
      "     41.3  0.0  MLOT    10.43s\n",
      "A->D .\n",
      "     39.5  0.0  MLOT     6.60s\n",
      "A->W .\n",
      "     48.1  0.0  MLOT     7.27s\n",
      "C->A .\n",
      "     55.5  0.0  MLOT    11.80s\n",
      "C->D .\n",
      "     52.9  0.0  MLOT     8.10s\n",
      "C->W .\n",
      "     54.9  0.0  MLOT     9.20s\n",
      "D->A .\n",
      "     40.0  0.0  MLOT     4.63s\n",
      "D->C .\n",
      "     33.7  0.0  MLOT     4.77s\n",
      "D->W .\n",
      "     85.4  0.0  MLOT     4.06s\n",
      "W->A .\n",
      "     40.1  0.0  MLOT     6.04s\n",
      "W->C .\n",
      "     36.1  0.0  MLOT     5.64s\n",
      "W->D .\n",
      "     75.2  0.0  MLOT     4.43s\n",
      "\n",
      "Mean results and total time\n",
      "     50.2  0.0  MLOT    82.97s\n"
     ]
    }
   ],
   "source": [
    "!python main.py -n --svd_solver full"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Cross validation run"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This code need to be run for each feature, each dataset and each method. This can easily be parallelized. The code bellow launch only one dataset for one method."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Namespace(SA=True, adaptationAlgoUsed='MLOT', cross_val=True, cuda=False, d=70, detach=False, featuresToUse='surf', lr=0.001, margin=10, max_inner_iter_grad=1, max_iter=10, new_space=True, numberIteration=10, pickle_name='test', reg_cl=0.1, reg_e=0.1, reg_l=1, reg_pca=1, rule='median', save_pickle=False, specific_comparaison='None', svd_solver='auto', test_exp=False, time_cross_val=48.0, type_classifier='1NN', verbose=False, which_dataset=0)\n",
      "A->C \n",
      "Train :  MLOT\n",
      "0 : result 47.32985386221294 target result 37.702582368655385 d 40 max_iter 5 reg_e 0.3 reg_l 100 reg_cl 0.1 reg_pca -1 lr -1 margin 0.1\n"
     ]
    }
   ],
   "source": [
    "!python main.py -n -c -sa --adaptationAlgoUsed MLOT --featuresToUse surf --which_dataset 0 --time_cross_val 48 --pickle_name pickle"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Cross validation analysis"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This code can be run only if the cross validation file are already saved. IMPORTANT : The -r parameter can solve some bug by deleting wrong run of the cross validation and reformating the pickle file.\n",
    "The parameter \"rule\" is used to chose how we find the best hyperparameter set. The easiest way is to use max which take the best set of hyperparameter. This is not the only solution, for example, \"mean\" is also implemented which take the best hyperparameter on average. \"precentile\" is also implemented."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "amazon caltech10 MLOT max surf_SA\n",
      "{'d': 50.0, 'reg_e': 0.05, 'reg_cl': 0.05, 'reg_l': 0.01, 'max_iter': 6.0, 'margin': 0.001, 'lr': 0.001, 'reg_pca': 1}\n",
      "A->C ..........\n",
      "     41.9  0.4  MLOT   560.41s\n",
      "\n",
      "Mean results and total time\n",
      "     41.9  0.4  MLOT   560.41s\n",
      "amazon dslr MLOT max surf_SA\n",
      "{'d': 80.0, 'reg_e': 0.05, 'reg_cl': 0.1, 'reg_l': 1.0, 'max_iter': 2.0, 'margin': 0.1, 'lr': 0.001, 'reg_pca': 1}\n",
      "A->D ..........\n",
      "     40.8  1.5  MLOT    63.85s\n",
      "\n",
      "Mean results and total time\n",
      "     40.8  1.5  MLOT    63.85s\n",
      "amazon webcam MLOT max surf_SA\n",
      "{'d': 100.0, 'reg_e': 0.05, 'reg_cl': 0.3, 'reg_l': 0.01, 'max_iter': 5.0, 'margin': 0.01, 'lr': 0.001, 'reg_pca': 1}\n",
      "A->W ..........\n",
      "     41.7  1.0  MLOT   214.64s\n",
      "\n",
      "Mean results and total time\n",
      "     41.7  1.0  MLOT   214.64s\n",
      "caltech10 amazon MLOT max surf_SA\n",
      "{'d': 90.0, 'reg_e': 0.05, 'reg_cl': 0.05, 'reg_l': 1.0, 'max_iter': 5.0, 'margin': 0.001, 'lr': 0.001, 'reg_pca': 1}\n",
      "C->A ..........\n",
      "     51.1  0.4  MLOT   397.37s\n",
      "\n",
      "Mean results and total time\n",
      "     51.1  0.4  MLOT   397.37s\n",
      "caltech10 dslr MLOT max surf_SA\n",
      "{'d': 90.0, 'reg_e': 0.1, 'reg_cl': 0.03, 'reg_l': 0.001, 'max_iter': 18.0, 'margin': 10.0, 'lr': 0.001, 'reg_pca': 1}\n",
      "C->D ..........\n",
      "     52.7  2.4  MLOT   808.35s\n",
      "\n",
      "Mean results and total time\n",
      "     52.7  2.4  MLOT   808.35s\n",
      "caltech10 webcam MLOT max surf_SA\n",
      "{'d': 90.0, 'reg_e': 0.05, 'reg_cl': 0.05, 'reg_l': 1.0, 'max_iter': 5.0, 'margin': 0.001, 'lr': 0.001, 'reg_pca': 1}\n",
      "C->W ..........\n",
      "     46.6  1.1  MLOT   241.26s\n",
      "\n",
      "Mean results and total time\n",
      "     46.6  1.1  MLOT   241.26s\n",
      "dslr amazon MLOT max surf_SA\n",
      "{'d': 80.0, 'reg_e': 0.05, 'reg_cl': 0.1, 'reg_l': 1.0, 'max_iter': 2.0, 'margin': 0.1, 'lr': 0.001, 'reg_pca': 1}\n",
      "D->A ..........\n",
      "     38.1  0.6  MLOT    49.74s\n",
      "\n",
      "Mean results and total time\n",
      "     38.1  0.6  MLOT    49.74s\n",
      "dslr caltech10 MLOT max surf_SA\n",
      "{'d': 80.0, 'reg_e': 0.05, 'reg_cl': 0.1, 'reg_l': 1.0, 'max_iter': 2.0, 'margin': 0.1, 'lr': 0.001, 'reg_pca': 1}\n",
      "D->C ..........\n",
      "     34.4  0.5  MLOT    51.00s\n",
      "\n",
      "Mean results and total time\n",
      "     34.4  0.5  MLOT    51.00s\n",
      "dslr webcam MLOT max surf_SA\n",
      "{'d': 80.0, 'reg_e': 0.05, 'reg_cl': 0.01, 'reg_l': 1.0, 'max_iter': 6.0, 'margin': 5.0, 'lr': 0.001, 'reg_pca': 1}\n",
      "D->W ..........\n",
      "     87.6  0.9  MLOT   126.30s\n",
      "\n",
      "Mean results and total time\n",
      "     87.6  0.9  MLOT   126.30s\n",
      "webcam amazon MLOT max surf_SA\n",
      "{'d': 80.0, 'reg_e': 0.05, 'reg_cl': 0.1, 'reg_l': 1.0, 'max_iter': 2.0, 'margin': 0.1, 'lr': 0.001, 'reg_pca': 1}\n",
      "W->A ..........\n",
      "     37.9  1.0  MLOT    70.96s\n",
      "\n",
      "Mean results and total time\n",
      "     37.9  1.0  MLOT    70.96s\n",
      "webcam caltech10 MLOT max surf_SA\n",
      "{'d': 80.0, 'reg_e': 0.05, 'reg_cl': 0.7, 'reg_l': 100.0, 'max_iter': 2.0, 'margin': 5.0, 'lr': 0.001, 'reg_pca': 1}\n",
      "W->C ..........\n",
      "     33.0  0.6  MLOT    69.69s\n",
      "\n",
      "Mean results and total time\n",
      "     33.0  0.6  MLOT    69.69s\n",
      "webcam dslr MLOT max surf_SA\n",
      "{'d': 90.0, 'reg_e': 0.1, 'reg_cl': 0.03, 'reg_l': 0.1, 'max_iter': 3.0, 'margin': 0.01, 'lr': 0.001, 'reg_pca': 1}\n",
      "W->D ..........\n",
      "     90.6  0.8  MLOT    49.57s\n",
      "\n",
      "Mean results and total time\n",
      "     90.6  0.8  MLOT    49.57s\n",
      "A->C ..........\n",
      "     41.8  0.7  MLOT   342.15s\n",
      "A->D ..........\n",
      "     39.9  0.6  MLOT   141.93s\n",
      "A->W ..........\n",
      "     40.6  1.3  MLOT   193.36s\n",
      "C->A ..........\n",
      "     51.5  0.9  MLOT   407.76s\n",
      "C->D ..........\n",
      "     48.1  1.3  MLOT   185.73s\n",
      "C->W ..........\n",
      "     45.3  1.1  MLOT   191.27s\n",
      "D->A ..........\n",
      "     38.4  1.0  MLOT   101.50s\n",
      "D->C ..........\n",
      "     35.0  0.5  MLOT   103.73s\n",
      "D->W ..........\n",
      "     85.1  1.2  MLOT    84.36s\n",
      "W->A ..........\n",
      "     38.3  0.9  MLOT   120.31s\n",
      "W->C ..........\n",
      "     33.4  0.9  MLOT   124.01s\n",
      "W->D ..........\n",
      "     81.1  0.6  MLOT    83.60s\n",
      "\n",
      "Mean results and total time\n",
      "     48.2  0.9  MLOT   2079.70s\n",
      "pickle_latex/testMLOTmaxsurf_SA.pickle\n"
     ]
    }
   ],
   "source": [
    "!python main_cross_val.py -l -la --adaptationAlgoUsed MLOT --featuresToUse surf_SA --save_pickle pickle_latex/ --rule max"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Latex result of the cross validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import recreate_table"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\\newpage Feature : surf_SA Rule : max Average : False \\newline\n",
      "\\begin{tabular}{lllllllllll}\n",
      "\\toprule\n",
      "         dataset &    NA &  LMNN &            SA & CORAL &            TCA &             OT &           OTDA &      OTDA_pca & JDOTSVMe &                   MLOT \\\\\n",
      "\\midrule\n",
      " A$\\rightarrow$C &  26.0 &  40.3 &  40.2$\\pm$0.2 &  25.4 &           40.0 &           33.9 &           40.2 &  39.4$\\pm$0.5 &     39.9 &  \\textbf{42.3}$\\pm$0.6 \\\\\n",
      " A$\\rightarrow$D &  25.5 &  36.9 &  39.3$\\pm$2.5 &  26.8 &           31.8 &           30.6 &           40.1 &  39.6$\\pm$1.1 &     37.6 &  \\textbf{40.8}$\\pm$0.8 \\\\\n",
      " A$\\rightarrow$W &  29.8 &  38.0 &  39.9$\\pm$1.3 &  26.8 &  \\textbf{41.7} &           32.5 &           37.3 &  39.8$\\pm$0.9 &     38.0 &           41.3$\\pm$1.1 \\\\\n",
      " C$\\rightarrow$A &  23.7 &  46.0 &  41.3$\\pm$1.1 &  23.6 &           39.8 &           41.0 &  \\textbf{52.7} &  48.5$\\pm$0.7 &     48.1 &           51.5$\\pm$0.8 \\\\\n",
      " C$\\rightarrow$D &  25.5 &  45.9 &  45.4$\\pm$1.2 &  26.1 &           44.6 &           36.9 &           47.8 &  51.4$\\pm$1.4 &     49.7 &  \\textbf{52.2}$\\pm$1.3 \\\\\n",
      " C$\\rightarrow$W &  25.8 &  41.7 &  36.6$\\pm$1.1 &  23.7 &           36.9 &           28.1 &  \\textbf{46.4} &  45.8$\\pm$1.4 &     43.4 &           45.9$\\pm$0.8 \\\\\n",
      " D$\\rightarrow$A &  28.5 &  31.1 &  35.4$\\pm$1.0 &  28.8 &           32.9 &           29.3 &           32.4 &  37.8$\\pm$1.0 &     32.8 &  \\textbf{37.8}$\\pm$0.7 \\\\\n",
      " D$\\rightarrow$C &  26.3 &  30.7 &  32.3$\\pm$0.6 &  30.0 &           31.5 &           31.7 &           32.0 &  33.5$\\pm$0.7 &     31.7 &  \\textbf{34.4}$\\pm$0.5 \\\\\n",
      " D$\\rightarrow$W &  63.4 &  77.3 &  88.5$\\pm$1.1 &  84.4 &           84.7 &  \\textbf{88.8} &           88.8 &  87.5$\\pm$1.2 &     82.7 &           87.8$\\pm$0.7 \\\\\n",
      " W$\\rightarrow$A &  23.0 &  32.3 &  32.6$\\pm$0.5 &  26.2 &           29.4 &           34.1 &           33.7 &  37.6$\\pm$0.6 &     37.6 &  \\textbf{38.0}$\\pm$0.8 \\\\\n",
      " W$\\rightarrow$C &  19.9 &  30.4 &  29.0$\\pm$0.6 &  22.6 &           29.2 &           30.1 &  \\textbf{34.1} &  33.3$\\pm$0.5 &     33.1 &           33.2$\\pm$0.6 \\\\\n",
      " W$\\rightarrow$D &  59.2 &  86.6 &  89.5$\\pm$1.0 &  84.1 &           91.7 &           89.2 &  \\textbf{92.4} &  91.8$\\pm$1.2 &     89.8 &           90.8$\\pm$0.8 \\\\\n",
      "             AVG &  31.4 &  44.8 &  45.8$\\pm$1.0 &  35.7 &           44.5 &           42.2 &           48.2 &  48.8$\\pm$0.9 &     47.0 &  \\textbf{49.7}$\\pm$0.8 \\\\\n",
      "\\bottomrule\n",
      "\\end{tabular}\n",
      "\n"
     ]
    }
   ],
   "source": [
    "recreate_table.creat_latex_tables(\"NA,LMNN,SA,CORAL,TCA,OT,OTDA,OTDA_pca,JDOTSVMe,MLOT\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\\newpage Feature : decaf6_SA Rule : max Average : False \\newline\n",
      "\\begin{tabular}{lllllllllll}\n",
      "\\toprule\n",
      "         dataset &    NA &           LMNN &                     SA &          CORAL &            TCA &    OT &  OTDA &      OTDA_pca &       JDOTSVMe &                   MLOT \\\\\n",
      "\\midrule\n",
      " A$\\rightarrow$C &  71.7 &           80.6 &           81.1$\\pm$0.3 &           72.2 &           77.7 &  82.2 &  80.5 &  79.7$\\pm$0.3 &  \\textbf{83.1} &           81.6$\\pm$0.4 \\\\\n",
      " A$\\rightarrow$D &  73.9 &           81.5 &           83.4$\\pm$0.7 &           72.0 &  \\textbf{84.1} &  77.7 &  76.4 &  73.9$\\pm$1.3 &           70.7 &           75.3$\\pm$0.8 \\\\\n",
      " A$\\rightarrow$W &  68.1 &           72.5 &           74.7$\\pm$1.0 &           64.1 &           71.9 &  71.2 &  71.9 &  75.0$\\pm$0.6 &  \\textbf{76.9} &           72.1$\\pm$0.3 \\\\\n",
      " C$\\rightarrow$A &  87.3 &  \\textbf{91.4} &           87.9$\\pm$0.4 &           81.3 &           89.1 &  90.4 &  86.8 &  87.6$\\pm$0.4 &           90.5 &           90.8$\\pm$0.1 \\\\\n",
      " C$\\rightarrow$D &  79.6 &           79.0 &  \\textbf{87.1}$\\pm$0.7 &           80.3 &           82.8 &  82.8 &  80.9 &  80.1$\\pm$0.5 &           85.4 &           83.3$\\pm$0.7 \\\\\n",
      " C$\\rightarrow$W &  72.5 &           78.0 &  \\textbf{85.9}$\\pm$0.7 &           68.8 &           79.7 &  78.0 &  80.3 &  78.5$\\pm$0.4 &           76.9 &           79.7$\\pm$0.3 \\\\\n",
      " D$\\rightarrow$A &  49.9 &           70.1 &           84.3$\\pm$0.5 &           78.0 &  \\textbf{86.0} &  84.4 &  84.2 &  82.6$\\pm$0.4 &           64.9 &           85.0$\\pm$0.3 \\\\\n",
      " D$\\rightarrow$C &  42.0 &           63.7 &           72.2$\\pm$0.4 &           71.8 &           75.2 &  81.2 &  80.9 &  81.0$\\pm$0.3 &           64.8 &  \\textbf{83.2}$\\pm$0.3 \\\\\n",
      " D$\\rightarrow$W &  91.5 &           95.6 &           99.2$\\pm$0.2 &  \\textbf{99.3} &           99.0 &  99.3 &  99.3 &  98.9$\\pm$0.3 &           98.6 &           98.6$\\pm$0.1 \\\\\n",
      " W$\\rightarrow$A &  62.5 &           73.4 &           76.1$\\pm$0.6 &           69.7 &           81.8 &  81.5 &  81.3 &  77.4$\\pm$0.3 &           64.4 &  \\textbf{85.8}$\\pm$0.4 \\\\\n",
      " W$\\rightarrow$C &  55.3 &           67.0 &           72.7$\\pm$0.4 &           69.1 &           73.6 &  78.6 &  75.7 &  78.0$\\pm$0.4 &           62.0 &  \\textbf{81.1}$\\pm$0.2 \\\\\n",
      " W$\\rightarrow$D &  98.1 &  \\textbf{100.} &           100.$\\pm$0.0 &           100. &           100. &  100. &  100. &  99.0$\\pm$0.3 &           100. &           99.4$\\pm$0.0 \\\\\n",
      "             AVG &  71.0 &           79.4 &           83.7$\\pm$0.5 &           77.2 &           83.4 &  83.9 &  83.2 &  82.6$\\pm$0.5 &           78.2 &  \\textbf{84.7}$\\pm$0.3 \\\\\n",
      "\\bottomrule\n",
      "\\end{tabular}\n",
      "\n"
     ]
    }
   ],
   "source": [
    "recreate_table.creat_latex_tables(\"NA,LMNN,SA,CORAL,TCA,OT,OTDA,OTDA_pca,JDOTSVMe,MLOT\",\n",
    "                                  features=\"decaf6_SA\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\\newpage Feature : office31fc7_SA Rule : max Average : False \\newline\n",
      "\\begin{tabular}{lllllllllll}\n",
      "\\toprule\n",
      "         dataset &    NA &           LMNN &                     SA & CORAL &   TCA &             OT &           OTDA &      OTDA_pca & JDOTSVMe &                   MLOT \\\\\n",
      "\\midrule\n",
      " A$\\rightarrow$D &  59.6 &           50.6 &  \\textbf{60.6}$\\pm$0.5 &  55.6 &  58.0 &           56.8 &           55.4 &  53.6$\\pm$0.0 &     58.6 &           54.7$\\pm$0.1 \\\\\n",
      " A$\\rightarrow$W &  54.0 &           51.8 &  \\textbf{56.3}$\\pm$0.2 &  56.1 &  52.7 &           49.2 &           50.8 &  53.3$\\pm$0.1 &     51.9 &           54.1$\\pm$0.2 \\\\\n",
      " D$\\rightarrow$A &  42.4 &           48.0 &           45.0$\\pm$0.1 &  42.2 &  44.0 &           48.8 &           48.1 &  48.3$\\pm$0.1 &     45.7 &  \\textbf{50.4}$\\pm$0.0 \\\\\n",
      " D$\\rightarrow$W &  90.9 &           94.7 &           93.2$\\pm$0.1 &  91.7 &  91.7 &           92.1 &  \\textbf{95.0} &  94.0$\\pm$0.1 &     90.9 &           92.6$\\pm$0.1 \\\\\n",
      " W$\\rightarrow$A &  40.8 &           43.8 &           45.0$\\pm$0.1 &  41.0 &  42.5 &  \\textbf{48.3} &           46.9 &  46.1$\\pm$0.1 &     44.4 &           47.6$\\pm$0.1 \\\\\n",
      " W$\\rightarrow$D &  97.8 &  \\textbf{99.0} &           98.6$\\pm$0.0 &  98.0 &  95.6 &           96.8 &           95.4 &  96.0$\\pm$0.0 &     94.6 &           97.7$\\pm$0.1 \\\\\n",
      "             AVG &  64.3 &           64.7 &  \\textbf{66.5}$\\pm$0.2 &  64.1 &  64.1 &           65.3 &           65.3 &  65.2$\\pm$0.1 &     64.4 &           66.2$\\pm$0.1 \\\\\n",
      "\\bottomrule\n",
      "\\end{tabular}\n",
      "\n"
     ]
    }
   ],
   "source": [
    "recreate_table.creat_latex_tables(\"NA,LMNN,SA,CORAL,TCA,OT,OTDA,OTDA_pca,JDOTSVMe,MLOT\",\n",
    "                                  features=\"office31fc7_SA\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AVG & 53.8 & 62.6 & 65.1 $\\pm$ 0.6 & 58.0 & 64.0 & 63.5 & 65.6 & 65.6 $\\pm$ 0.6 & 63.0 & 67.0 $\\pm$ 0.5 & \\\\\n"
     ]
    }
   ],
   "source": [
    "recreate_table.average_result() # The value are hard coded in this function"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Reproduce Figure 1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " This code can be run alone and doesn't need any other file."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import toy_dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "toy_dataset.main(algoName=\"NA\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "toy_dataset.main(algoName=\"OTDA\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAV0AAADnCAYAAAC9roUQAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nOy9aZSl11nv99v7Hc98au551iwPklu2sYxlbItrCJMxkHBtyVyyrny9uCa5ZtkOU1ZInEUIAZPgxWUp+AJSQgy+EPBdMQZdMDYSg615dKvVc1d3zVVnfMe9dz68p1vd6knqoaq6av++dJ13fOpUn/953mc/gzDGYLFYLJblQa60ARaLxbKesKJrsVgsy4gVXYvFYllGrOhaLBbLMmJF12KxWJYR9xL7bWqD5brkgx/8IF/72tdW2gzL+kVcaIf1dC1rkrm5uZU2wWI5L1Z0LRaLZRmxomuxWCzLiBVdi8ViWUas6FosFssyYkXXYrFYlhEruhaLxbKMWNG1WCyWZcSKrsVisSwjyya6P/dzf7Rct7Jc5+zbt4+PfOQj3HLLLTQaDcrlMjfffDOf+tSnOHny5EqbZ7FcEZcqA75qPPvsseW6leU65/jx45w8eZIPfehDbNmyBdd1ee6553jwwQf50pe+xNNPP834+PhKm2mxXBbLIrpxki3HbSxrhPe///28//3vP2f7e97zHn7iJ36CP/iDP+Azn/nMClhmsVw51zy8oLXm3/zWS9f6NpZ1wPbt2wFYXFxcYUsslsvnmnm6Dz30KA8//NhZ2+6999dO/3zffXdz//3vvla3t6wB4jim2+0SxzEvvvgin/3sZwH4/u///hW2zGK5fMQlBlNeldaOh0/0+Ncf+wJv+vAP8Zv/5parcUnLOuALX/gCn/zkJ0+/3rFjB5/73Of4yEc+ct7jH3zwQR588EEAZmdnOXLkyLLYabGchwu2dlyWmO6OTRUAnj/U5eBkl12bq8txW8t1zo/8yI9w88030+12eeqpp/jKV77C7OzsBY9/4IEHeOCBBwDYu3fvcplpsbwhli17YevODRxShj97dJqf+4kKQlzwi8BiAWDLli1s2bIFKAT4wx/+MHfddRdRFPHzP//zK2ydxXJ5LFue7u984aNsGg751r42k3Pxct3WsoZ485vfzB133MHv/M7vrLQpFstls2yiG/oO731znYVuztefnOMSsWSL5bxEUcTCwsJKm2GxXDbLWgb8gbeN0yy7/ON3OpycT5bz1pbriKmpqfNu//rXv87zzz/PO9/5zmW2yGK5eixbTBdg+0SJW7dVePpwl2+/tMgPvXuDje1azuETn/gEJ0+e5H3vex/bt28njmOeeOIJvvSlL1Gr1fiN3/iNlTbRYrlsllV0pRS8+01DvHi0z5MHu9x1a8KmkXA5TbBcB/zkT/4kf/iHf8jDDz/M7OwsQgi2b9/Oxz/+cT796U+zbdu2lTbRYrlsliVP90xa3Yxf+v2X6USKD79ngh94x7j1di1Xnb179/L444+vtBmW9cvqGcHeqHrctr1GnGoOTPY4YTMZLBbLOmJF+um+6/YhamWXyfmUZw52bCaDxWJZN6yI6N6yrcLGIZ9uT3F0OuLkvPV2LRbL+mBFRNdzJXfc0EApw1Iv45kDXevtWiyWdcGKjet5x80NKiWXXmyYnItslZrFYlkXrJjobhwJ2LExoNPPSTPD84d71tu1WCxrnhUTXSEEe29skiuDEYKphZgTtkrNYrGscVZ0GvAde2rUKx7tXkaeaV443EVr6+1aLJa1y4qKbjl0uXV7memllGrF4eRCYr1di8WypllR0QV4+81NpIAoMRiteeFwx3q7FotlzbLiortnU5mJoZDDMxGjDZ+ZpZTJOevtWiyWtcmKi66UgrfuqdHu5ggJUgi+c9TGdi0Wy9pkxUUXYO+NdcLA4dhswnDNZXop4/is9XYtFsvaY1WI7kjdZ8dEiaNTMY2qhyNhn/V2LRbLGmRViC7AnTfUyZRhvp0y3nCZbWccm7VVahaLZW2xakT3zTur1Cour0xGNAfe7v7jPevtWiyWNcWqEd3Ad7h1W4WT8wm5gg1DLjOtjKOz0UqbZrFYLFeNVSO6AHtvbOA4ggNTEc2qj+vA/mN96+1aLJY1w6oS3e0TIeNDPgdP9BACtgwHzHest2uxWNYOq0p0hRDcsbtON9Is9hT1iovrCl45HqGU9XYtFsv1z6oSXYA799TxPMnLx3sIKdgyErLQSTk6Y71di8Vy/bPqRLdeddm9MeTodAwIhqsOriM5eDIiz/VKm3cODz306EqbYLFYriNWnegCvO2GBlrD4ak+Rgo2jwYsdDKOzqy+vN2HH35spU2wWCzXEatSdG/dXqVScnhlsg9GMFp38Vaht9vppAB24oXFYnndrErRdRzB7TsqzCxltPs5GMGm0YBWL+fI9Orxdp881AGgPRBfy8ry4IMPsnfvXvbu3cvs7OxKm2OxnBdxCS9txVy4qYWE3/7zo7xlZ5U7bmzgCM2zB/vUyi73vLmJ667c98VDDz16wbDCfffdzf33v3uZLbK8lr179/L444+vtBmW9Yu44I7VKroAX/iLI3R6ig+9e5ySK5npZByfiXnTrhp7NpdX0jQAfu+rx/jjz/8Rjzzy2Qse89BDj1oRXgGs6FpWmAuK7qoML5zizj11urFiZiklN4bRuofrCo5Mxyse201TRTdSlzzOLrRZLJYzWdWie8fuOp4rOXCiR64hcGHjcEC7l3F4hWO7S70MpQx3v/+O8+5P04y/eGx6ma2yWCyrHXelDbgYpdDhxs1lXp7scdfNCm08hmseUwsJh6cjtk8EeK6zIrbNt1KUNvzgj77jrO3ni/fee++vATbea7FYVrnoAuy9scYLR7ocn4nxNjo0K4INwwHH5xIOnYy5cWtlReya7+Q4jqBROfstvP/+d3P//e/mz/5+ildORDzye39y0ZivxWJZX6zq8ALAns0VmlWXAydihACtBENVj8ATHJ2NSbNLx1WvNlmmafVyKqFD6J/raaep4uhswuYR/6LXsdVsFsv6Y9WLrpSCN+2ocmIxJYpyepmhVpZMNH36keLQ1PLHduMspx8pKoGL5567SHlsNiaKFbs3XdwLPxWGsOJrsawfVr3oArz95gYCw4GpotuYK6BWcQl8ydGZaNm93X6kSHJNueTgOeeK7v7JPp4n2TERct99d5/3GkenesW1+qnNcLBY1hHXhegO1322j5fYd7yPK6EdGxpll/GmR5JpDk4tbweyhU6OEFDy5Tmebj/OmZxPGG941MveOQtnDz30KPfe+2v81/d9AYAf/uHPA8Vim/V4LZa1z3UhugB33lCjGynm2ylJpgkcqJZdAs/h2MzyxXbzXNONc6QUVEsOjjz7LZycS+jHio0jPoF/7tt7//3v5g/+70/y3p/6MAD/4+f/NQD//j/8W5vZYLGsA64b0X3TziqBJ3npWB8poZcaaqHLWMMlzTQHTi6Ptxunil6kkUJQCiTiDEdXKc2x2RgpJZtGQpzzhB4AHnuxzZFnXgTgv/93/ycAn/jpL3Dvvb9mPV6LZY2z6lPGTuF7xeDKZw93+e7bm3QixZYRj3bkUglzjs/G7NpQIjhPNsHVJMk0aa4phw6eI5FnqG6nr5hrZQxVXZpV77znH5+NODzd5wd+5C5++6mXuOdjH+Ybf/inNq3MYlknXDeeLsBdNzdQynBoOkIIiBJDteQwUveXzduNUkWmwHMkriM4M7owvZgQpznDdZdKeO5bq7Xm2/tauFKwa7zEnrfdypbR4JrbbLFYVg/XlehuGw8Zbfg8f6hH2ZO0IsXQIIuhUnY4PhsTJ9cutquUJko0AkPgSaQUp8MLSaqYWkyRUjBU8c6bv3t0KubobMKuDSUOTsdsf/OtfPebGnz0o++6ZjZbLJbVxXUlukII3rq7ysxSSi/NyZUhyzXlwGGk6qO0uabebpQoerHCcSSBLwpPd6C6rX7OUi+nEjg0qh7ua+K5WaZ5/kgXCTQrLoemY3ZvCtkyVrYLaBbLOuK6El2AvTfWkVLw0tE+niNY6iuGa4W3Wwtdjs9dO283yQpP15WC0JM4UiClwBjD3FJKlmuqpaJazpFni+6R6YjJ+aJKbXIxAaO5Y3cd31uZ3hEWi2VluO5Et1b22L2xxHOHelQCUTzuGwg8SbPmoq+htxtnGqUF0gHHkbgDYe1GOQvdHNeFWsWhHDiIMxbY4kSx71gfRxbe+nwrZ/fGChPDRTz3zGMtFsva5roTXYC33VgjThXH51KkhKV+zkjNw/ckjXLh7Ubx1fV2ldIoBcpoAk+itcH3BqGFnqLTzyn7HuVAUg7P9l4PT0fMtlLqocNsK6Vadrhhc5lq+fwZDhaLZe1yXYrubdtrVEKHpw92qIUO3VhTCgSBJxmquxgDB072r+o940SjlEarYhFNCHCkJM817X6O1oZSIHEHWQ2n6PQzDp6M8V1BP9WkmWb7WMBY07cersWyDrkuRVdKwe07qqeb3RhjWGorhqoujpQMVR0m5xP6cX7V7hlnRb8FxxFnZS60ejmtfo7vSspBEes9U3SPzsS0uglGFJVsI42AreMlapXrJkXaYrFcRa5L0QV4xy0NjDG8cKRH6AlaUU510ICmWfHACA6cuHqx3TQvwgtQxI+lKHJ02/2cKFZ4rqTku/hniO58O+X4bIIrJd1IIYVg61hAs+rhOtftW2+xWK6A6/aTPzEUsHk04KkDbZpVj0wZen3FcM1DSMFwzb1q3q5SmiwvhNd1iqwFISBJNHGqT4cWoGiCI4RAa83x2Zh2XxEPsh42DAeMN3zq5cLLvcRQUIvFsga5bkUX4K27ayx2cmaX0tPpY/WKiysFQ1UHgeCVySv3duOkGIKpNASeAAFSQCcu8nY9V1AJXaQUeIPR8NOLGTOLKWhDLykanm8ZDaiWC2/YYrGsT67rT/+de+p4juTx/R0aFUk/UaSpYqjmohGM1F1OLCT0oivzdpNcY4whV4bgdF6toRer095vKSiKJVxHkOeakwsx/Tinm+SkmWHjkE+z6lEvO0hpF9AslvXKdS26YeBw09YyLx3tEbhOkT7WUzQqLo4QDNc8pLhybzfJFI4jyJQh9CVaQ5watAatzaDkV+C7heieWEhY6OREmSZNddH7dyigHDoEvrRhBYtlHXNdiy4UFWpZrnnhaJ+KL2lHCmMMzapLrg1jTY+Tiwnd/uV5u6fiubkyGGMIPIEx0E8USmtAUClJ1CCVLM81s0sp3SinEykQgtGGz0jdpxI4Kza92GKxrA6ue9Hds7lMo+LyxP42Q3UPYwztrqJZdRGi6HMgxeVnMiRpEc/NlUFKge9Jkiwnyw0GgxBiUIEGjiM4uZDSiYpiCWMo5rmNeLiupBScK7g2V9diWV9c96IrhOAtu6pMzid0IkXgCpb6+UBwi6yGiaHL93bjTCMEZLnBH/TPbUcKIUDlUAmLbY4UxKlisZux1MmJUk3Jd2hUfOolj0og7QKaxWK5/kUX4O03NREYHt/XYajmkeaGKFY0K4W3O1wrvN39J954lVqSKQKvWKQrhRJtDP1Y4QCZMlTLDtoUXu7UQkaUKBa7Ca6AciAZb3r4nqAUXB8LaHZqhcVybVkTojtU99g+EfLMwTblQOJIWOwqPE9SL7nEqWHzqMfUQkqnn73u6yqlyPIiPSzNDZXAod3LMAiUKCrjqoFDrjT9WJFkipmllFxDrVL0YaiXHUr++Ue1r0bsZGKL5dqyJkQX4I7dxeDKlycjGmWHXqJIkpzhmocBmlUPRwpeeQOx3SQtsgyEEOTKUAkdFrs5roA0M5T8ohwYA7OtlDjTtPqKcujgu5KxIR/XKWapvbaF42rMYOh0k5U2wWJZ86wZ0X3r7jqhL3lif5tm1UdKaPULb7dWcogSfdrbbfden7eb5EU8N9cGMAQetHsK3xP0Y8VI3SdThqVuRprDXCvDGEU1dCmHRX/fUuCsei83ijP++vFZvvCVYyttyhXx4IMPsnfvXvbu3cvs7OxKm2OxnJc103XFdSW3bqvwzMEuaa6L9LG+YrShGa55dCLFcNXj5HzG/smIt9146baKSaYJfEmnr/AcST8pvFMpirSxRsWh3c+Za2fEiaIb5TQrPp4DIzUXBJQD54yCildZDVkLcZLz6//H3/LNv37qrO333vtrANx3393X1VSLBx54gAceeACAvXv3rrA1Fsv5WTOiC3DXTQ2efKXDk690uHNPje58Qrub06z7VEKHfqrYPOpxeLrwduuVCwuvUoo0MzSrDlNxiudK2v2MwJdEqSYMJJ4rmV5MUcow38lwJJRDh3IoKQUugQeuI1fdAlqU5Dyxv8XzBzp4W3fzk5+6hXfd0uCm7TW+93v/VzuZ2GK5hqwp0d02UWK86fPk/g7f/aYhPJnSihT1qma45nJsVjFSC5icy3h5ss/eGxsXvNapeK4jBHGq8RxBlBoCV9DqKkYbPv0kZ76dEWU5cWbYMuIDgqFqkS/cKAenG52vBnpRzrMH27xwqEsnVQxXXO66qcGN26rUbEN1i2VZWFOiC/DW3VX++okFjs1GDNc9phZTkkRTKrmEfk4v0WwbDzlwMmKpm9Gsnl9sklwXHqooCiNcaYBCQHMNI3WPw1MxudYstnPKgSDwHaoliSMEJV/iOOKiM9CWK8TQi3KeOdBm32SXXmwYa3i8eazKlg1lhmtnTy6+7767l8Umi2W9smYW0k6x98YGUgoe39emWnJwJCx0i6KIkZpLpgqv13MEr1wkbzfJNL4nSHNNliuULpqXd6Kc0BdoDPPtjF6Uk2Sa0ZqHlNCoeDiDjIWVXkDr9HL+4flF/vSxkzx9sEOj5HHPbQ2+69Yhdm6uMtbwzxkVfz3FcC2W65E15+lWSi57NpV57lCXH3znGLWSQ6uvSdOcSskl8DK6sWb7RIlXTvTP6+1qrUkzUyyURTlJZmhUJYEr6CeaZsVlZiGhG+VESZGtEPgujXJRhCEFlAIX312Z77R2L+eFwx32n+wRp5rRasBde0KG6sUgTN91aFbdVRdrtljWA2vO0wV42w1Vkkzz3OEuzaqHEIZ2rxj7MFzzSHPNyMDb3X/8XG/3VP/cwJP0I02Sasq+INeglEEIWGgXpb5KQbVS5OXWy4NG5kHR6NxZ5ukQS92Ux55b4C/+cYqXjvVoBB7vua3BPW9pMjYUDvpEuAzVrOBaLCvFmvN0AW7ZVqVamuPJ/R3u2FOn7AtakWK4oamVXeZaRVOanRtLvHy8z2InZajmnz4/zTVCCDwXFrsZ2sBQzeU7xyJ8T9Lua+a72cCjdQhdSb3qYCh6MJR9Z1kX0BY7Kc8f6nJ0to/KYbTms2OTz1izRDlwitHxylAruedMKrZYLMvLmhRdx5HcvqPCP32nw2I7o1n1OT6X0OnlNGo+wzWP6aWUiYbHYa/ot3vXza+KbjyI5+ZK0OrlhIGkGrosdXOU0uRakeVFb12BoF5xqZYc8gx8HzxPnp4g8VqMKVpEXo1FtPlWyotHuhyb66OVZKIZsH0ioFbxqJY8fFfQ7iuMgaGqZxvuWCyrgDX7KXz7TQ0whideaRdluY6g1ddoralXitBCK1LsmCgx185YaKfAq/Hc0JOkuabVyxmteSSZoRflaGNY7CgQ4HuCMJSM1Dy0MkhZ9Nv1XHFNMxPmllK++cwCf/XELEdnYiaaIe+8ucbtu6qMDZUYawa4jmCplyNlkWlhBddiWR2s2U/ixHDI5tGQp15pA9CsuUSpIk2L0EGz6hKnmommi+8J9g+mS6TZq/Hc+VZKpgzjQwEzrZROPyfONFKAJ0EpGK56lEMHbYr5aL7j4F2jWO7MYso3n5nnkafmOT4Xs3kk5LturnPLtgrNekCz6tOsuvRjTbufE3iS4VqRTbHS2O5lFkvBmhVdgDt2V1jq5hw4EVEruUX3sV6RPnZqpM9SX7NzosRCJ2O+lRIPRNn3BCcXUwJX0ihLpuZjeolCGDjVqqbkOww3PHJl8ByB50DgO7hXMWvBGMPUQsw3np7nb59eYHI+YetYwHtub3Dj5hLVik+t7DFc8wg8yVI3J0oVldAZLCKuvOCC7V5mub751V/9VX78x3+cXbt2IYRgx44dl32tNRnTPcWdNzT4qycWeWJ/mz2by6fTx0YzhecVaVPznYwtIz6HposOZDsmQvzBSJ75Vkqj6qCBIzMxSaqolYt+C9KRjNQdyp6D0gbPKzzdq5Wba4zh5HzCy8d6TC9lGKHZOl5i13iI44ChiBvXyg6eK8mVZqmbo3XxhfLa/FuLxXL5/MIv/ALDw8PceeedLC0tXdG11rToBr7DzVvLvHikS5woGlWXdpTQ6SmGmw5DNZfFbkarr9i1ocSLR7oELuzYUGahU3QO2zTiMbOYMr2QEgYu4SB7YddGj+GaT6aL0mDHKQZTvt7Y6YU8UGMMJ+YTXj7aY7adYTBsHQ/Ys7GEIwWZAiGLzmmnxv/EqaLdK6ZZDNXcCy7iLSfGGOYWEw5ORxydti0jLdc3Bw4cYNeuXQDcfvvtdLvdy77WmhZdgLfdUOO5Q12ePtDhnbc2Cb2UpX5Os+4ipaRR8VjsZmwbDdh3THB4OubmbVVOTCXFyJ+ay/OHenQixZsmQqJMkyvF5pEQ1xFobfB9iRSGwHcu+3HeGMPxuYT9x7vMt3LAsH0iYOdEmdCTJMqgjKAcSqqlV6dQdKOcXlx0QVvJgockUcy2YybnEo5MJZxYSHj+H5/lyDMvnT7mVPcyeLWD2UMPPWqr4CyrnlOCezVY86K7Z3OFZtXlqYHoNqs+J+YTen1FrSoZqros9TJakWLjsMczB/vMLMX0oqJvbi9SHJqOaFZdxoYCXjrapVp22TgaMttKTw++dB1xWRVoWhdi+8pkIbZSwI6JkG0TAdXAJVGaOC+6lZ0KJUAh0q1eUYJc8h1q5csX/DeKUppuP2e2nTK1mHJiLmF6KaUXqaLZe8lhtO5x30fuZvdnPsDmsdIFu5c9/PBjVnQt64o1L7rF4Moa33h2kZnFhNGGh+cIlvqKSrlY9GqUPVq9jJGaR7Xk8MyhPo2KQzWUHJ6OaHVzbtteIU4V3UixZ3MZMGSZJvSKScChL1/3ApoQAqX0wLPtsdjJcSXs3BiyZTSgVnJQWtDPiiY71ZJzVlGDUoalXka+DAUPWmuiWNGNFbOtjOnFpFh07GQDkRWUA8FI3eOtu2rsmCgxPhQQ+Bd/L+aXEibn42tmt8WyWlnzogvw9psbfOPZRb69r81/8c4xmlWX2VZGlmmCoPB2FzoJsy3Nnk0lHn+5jTEeJc/hleMRow2PkbrLgRMJldBhx0TIYifH94t5bNqY171wlWWK43MJB09ELPVyfFewe1PIxuGAesVBCkkv0WhdFF+cGUqAIqWt1cuvScHDqRzlfpzT7ivmWxlznZR2X9HuZ8RJMUHDdSWjNZ9btvpsHg0YawZUQ+e8qWlxnLPQKSZ1/Kd/mOZv//Jxnv+n58865nptmm6xXA7rQnSbVY9dG0s8e7DDB+8aoVZyWehkLPZyNgQunicJPMnMYsbuDT5SCo7NxPiuoJ/l3LajwVJfEaWKTaMB5cBhtpUx2vBADNLFLpG1kOeaI9Mx+4936cbFhOHdm0M2DoXUyhLPceinmjTTOA40q+45gtqPFZ0ox3UEzcqV5d8aY8hzQ5zl9CLFUlex0Elp9RSdOCdJNVluit9PSsYaAc2qy2jdZ6ThUSu5lAKJEGJwLU2nr5hrpcy2UqYWUmZbGUu9nDTTbHvLLTx3uMeNd72Juz9wJxPDAVtGAx64/wu2abplXbEuRBfgjj01/vTvZ/jOsT637ahSCSSdSDOaa1xXUg0cDDDXUWwZDXj0hYg4ydm+ocJ40+UfXupSKzlsHPLpJ5pcF13IMgW1kkTK83ucWa44PB1z+GRMu5/hu2bg2ZaoliSB5xCnmlZSNOSplpzTYnYKYwydgegHnqRRcd9w/DbPNXGqSFJNOyrEu9XLaXVzokSjtEEIg0Tie5Jaw6VWKsqbm1WPetmlHAgcKYlTxUInYWEyZ6GdMtvOWGjndONCYLWB0HOoVxxu2lxmw7DP5n/xAcabwUX7C1ss64F1I7pv3lnjq/88zxP729y2o0qz5tFNin4MQw0fIwTVUDK1lLFlxGW2lZCkhvffWWK+rUgyzQ2bQ6olj/l2Sjl0cIRAYQj9c9/GLFccmoo5PBXRi3MCz+HGLaVB3LgYWKm0oR0plILAl+d9RNe6WDBLc00ldKiWLv0nU0qTZIXX3E81rW5GlBZVap2eIlMGbQoPPfAkQzUP1xV4jqQcSAJPUAocSp7EoOlEipePx7T7iqVuxmIvI4o1mdIYDZXQpVp22DJWZcOIz0TTZ7jmn/PlYbFY1pHouq7klu1lnn6lS7efUy27BG5KK8qpVYo+CyXfAZMXeaW6SM9KUsVCJ6dWkmwYCknzoqXj1tGAVEHoi7MW0JJ04NlORfSTnJLvcMPmEiP1YoHMc0BKQS8p4qcXCiUAr6vg4VQcNskUSWbo9HN6iSoW/fq6aD+pixaUpVBSr7i4jsR3QWmDlAKBAIry5iRTdPqGThTRiXK6kSq8Vw2IIvd5qOyzfdxloukz2vCpV1zKwYWb/FwMO6nCst5YN6ILcNdNdZ7c3+GJ/R3uecsQzarPyYWEVludFpXQE/z9y11Gag5B6LLveJ9a2WXjcECt7HB4Ois8xKCoAisPGqDHieLQVMSR6XhQhiu5aWuJZjWgGkpCXyKFIEoUcVYUElcGQyzP5w2er+DBGEOaaZJMkeaGONF0BpMr4lTTjzW50mQKMIZyqai689xigrHKDUmuidKMxU4RBshVsRCock2UFbFZEEhp8D2H0XpRZjxUcxmqutTKLuXg3BDI5WIXzizXAw8//DBHjhwBYHZ2ljRN+dznPgfA9u3bue+++173tdaV6G4bLzPe9Hn6QCG61ZLEcwQznQSBQIhC7Ba7GVtHA4ZrHs8e7FIJHMabPmlePLI3qy5Kg+cU5cIvHO5ybCYmHpQJ79hQplHxqAzE1pGSXBlafUWWa0qhe8HVfigKHrpRjkAQeoJOLyfJNb04J9DNEUUAACAASURBVE4NSabJck2SGZQyZLlGSIHnCELfYcgXSAn9xNDt53RjRS9WJLlBYnAdgSNBCAeExpUSz3OoVYoQR7nk0Ci7VEKXSsm5bC/WYlkrfPGLX+Qb3/jGWdt++Zd/GYB77rnHiu7FODW48uhMzLbxkFrJYXoxwZUGx5VML6XUyy4OAikMnitRuihO6PQVyhiqJYduP2Opl9LpF/PU6mXJro1lamWPki8pBRJ30G2sFxeeqJSGRsWlFJ77titVLHTNLmV0IoUYiOF0bkgyg9bF4l2SKbQqplhIWQhoJXQQUpBmiqVezsGpjF6syXON60gCX1IbhBakKDItXEdSCR0CT1IOCo+7EjpX1Yu1WNYKf/d3f3fVrrXuRHfvjQ3+5qlFHt/XZtt4SL3skmSKVqqLrAQl2LOxxLGZmLxluHVbmTgznJyPadRcpIGXjvTp9FNqZY+RusfuTSWqpVfF1pESKSFKimtCEUooBYWQKaVJ82IMUJIVIYMoU8wtpUSJIfSLqRVZlpPkhjQ3RfxVgDLgu+C5giQzLHVzOpGin6jB1GJJs+owMRRQLzv4XuHNO7JYNHNdSckvhNZ6sRbL8rPuRLdSctkx4fInX/6PPP6XJxgdHWHHmz9IO6uR5IadG0PaPUUvydHG5R23DPGt77R47nCb8UZAuxsjpM/OCZ9bd5apljzKvkspLGK2UhYj29tdXYigA6EnybVivq2JE4Uy4rToZrkmSjRL3Yw0U/ieQ5JL0lSjMRjAHRRHaG1QCuYTRSfSSAmeU4joeNOnWXEJA4kxRfMdKQcZCdaLtVhWDcIYc7H9F915PXLo0CG+613/gsWlcdL4XsLwENp8hQ/e91luu+VGtjRTFtlJzBD0D3P8mf+PJ549wuTkEZam9mF0yviGm/jMz3+Kf/WxjxapY7IQ2yK9S9Hu5eS6EFwpi5LfKNb0U0Wea5QGgyHPYbGnaPcyMmWohA5SFG+6IySOCxjIVCHkudJ4TpHSVSsX8dZg0NlMiGJaRckfCGy4vr3YvXv38vjjj6+0GWua4XqdxU7nrG0ukF/g+KFajYV2+4LXudi5Fzt/lXJBz2Zdia4xhttuezv79v0kWn/qjD3fAd6OED7GFL0yvWAYow1aPYDWO4GvAv84+HeOcvln+JVf+Rk+/olPkOZFvu1iOyXLizQsYwxaaXIMaZLxj998hO+8+BzVxihvf+8P44c1+nFOPy2mVNRKRXw1DBxcKciUHhgN0hFIISgHReFCyZfWi70EVnSvPUKIcwRCUIjGMLB4gfNeK56nrnPq3Aveb/CvC9QuIeCv577XGCu6AN/61rd43/s+Qq+3j3OHZvwC8C3g/wW+DfxL4Clg4xnHPAx8Cvgc8E5KpffzN//0LP3MI840jiMIPYkQhWcaJYajhw/xyz/7L+n3NhBH34MfvAL8FR/7d/8Ld3/PDzBa9xiqeoNcX0M3UkXRgSnyZ31PDOLBjvVi3wB79+7lZ3/2t2xK2mVyCV0AQEp5QdG9mICK11z/jYiuATwu7BG7QPY673uNuaDorquY7sGDB8my2zj/lKJ3An8O1IAvAT/H2YIL8FHgfwYeAn4LxC6+9p+/yVvf/l5cit663V5GlBVpXEpr/oef+2kW5/9bjPkkAGkC8DQP/dYH+KHvfSfjzR10I8VSLyXNDaVAUglcGhWPesWxXuwVsBrbRl72h/51nnfO9a+RyFypeOk8w6gcnbzxTnNFt+nzcz18QtaV6O7cuZM0/TagOVd4HwdOAJ8HJoEfPM8VBPAm4MeBJ0jiP6KfFFVjaW5wpAAEtZLDSM3n2P5v02sbjPm3r7nOW1H5x/j3v/N7/NQnf5Fq6DJS9xltFLmx5/Ni41Sf9Xr5vrAvD61fNdBccPsZP5uzjzGm2KZNcZQxg+2Ywc/FtlOvzWB23alzT71flxSHwf43LCKr7Q9woS/kS31Rv4EvchX3yaMO2dIsydzxN2DcuUQzx8ja82StuSu6zvXIRUX3gx/8IHNza+dNKT5YU8AOYPyMPTGwD9gJfJoiIvUksPm1VwCeA14EfLSe5P958HN4nocjBY5TZBpIURRaLC4uEEcd4K5zbMnzRb7+1Xn2P//35/y/N6/54ZyP9+nhmOa1m8852FzghTnfZnOe613i3LPtORdz0RPPf63LJYki0n7hOfV6Mxw9PEOj8dun94+MVBkZrryxi67npwtjMFoVHmmeYXSOUTlojdFFg6a95zlt72v+PR93f+D7wBjEoFHU6znn9R530X17L3WHSzM7O8vY2NhFj3niiSe+Zoz54Pn2rauYrtYax6kAdWAX8APAIeDPgN8Efgp4gEJ0vwj8J4qwAxTe8S9SxH3/BgDHGeN/+4OvsmHzFqQURSnxGZ7Zi0//A7/xy79CHD3Dax98pPPf8d7vW+InHvjF0/GsV3n11ZmfeXH+Q16zvRB8IV7f8eI1L8Rgv5Ti9DWKf4vXUpxx3Kn7CIE889gzzpWD/WdeEzj9xSQGaXZn3XdgpHPqZyiOEZxOyztl06v7xMC2Yt+Hvu89bBj/L/nrv/6MDcu8TnSeofod0tYsyfwJsvY8KupijAYjMEYBEpX2Ea7L9h/8xGXHdBde/GeCxghOWCEc2fiGYrqXjBdfbN9VeEJ5nYu0NqYLReB/+/adHDmiga1AF7iVYmFsw+CozUAf+K+A70GIvRhzC/B3wBjFQhvAS4Qlh3e8ZSee7w2E40whEdy+416++Ju/yHT8+xjz02dY8h0874t89tOPsHvPyFmiJs4QqVfF88x9l3p9ZeeuFYJBA6G19DtdbYxSRbigPU8yN0m6NIuKuqg8xQlCnKCKP7bldOwma8+R99q4lTrCcWiWQ0T/7Jisy+uLq5Y37kR6AUYrhiplRK9/yXOHruB3XU2sK9EF+OM//n2+67vehzEe8Mec/Wc2FIto+7nrrnv4/Of/M1/5ylf4/Od/jyz7XeDHBscvUC4/wGc+/d/wXW8avej9/uaRP+Oeez5IHH+Zbvd7CIJXEOI/8ru/+7/zve956zX6LS2WczHGoKIuaXuedP4E6eI0eb+NSmOkFxRe54YdeI0xBJBH3UKUW3Nk3UWE6+OEZWRQQQYhBx77KsKvIByHoDmBDEuoqItO+my99W2IfnReO4aqVdxKA+G49I/t4/DTj+ENb6T90j/hD2/Cb4ywZc9tLEVnC/oir0/QL3TMUK32Rt6ua8a6E913vOMdfPvb3+Rd7/pe0vTngV8BAiDBdf8ntm/3efbZNuVyGYC7776bLVu289nPfhzX/Q9AiTT9Oh/72L/il37p0hMPbr31Vo4e3ceXv/xlnnzyOTZvvpGPfvQFNm58bWaE5Wqz3ttGGmPQSUTanieZnyRdmCbvtVBJD+kGyLBMMLYVvzmOU6oi3YA8apMszqD7LfJ+IaBGCtzKEG61hluqI/0At1RHBCVMGoPOMSYn6yygkgiTpzz/5w8iXY/y1psxwhAObcIrVxGuj3CKFqXJ4gw66ROMbqE3+TLC8Qg37qZ74Ele+Mv/i8qWG5j5xpe562d+nVacnP69LuYRD1WrvPSXD+GGFWo33Il0PZywevqeV4MHHnjgis5fVzHdM5mamuLHfuynePLJp/H9W0nTF3jb2+7ky1/+fTZs2HDO8Z1Oh0ceeYQ0TXnve9973mMsq4f1Whyh0pi8s0g8P0kyf4K8s4SKe0jXQwYl3MYYQWMMp1LH8cIi+yNXqH6LvLtIHnWLRTOj0XmK4wc4pUbh4Xo+TlhFui4mTRCOSxZ3EQjcco08STBpn7TbJls6QX33HXjNCVTax62M4JZKp8M9JkvpHfsOMiijs5R49iiljTegjKZz6GmqEzvRKmf2H/6c8pabKG+5iWj2GJWtN+O6LllnCadcIWvNky6cZPy7fww3LNHe/yTRyUOUt95MaWI7GIUMykjXW+4/hY3pvpYNGzbw6KNf48CBAxw6dIhdu3ZddLZ9rVbjR3/0R5fRQovl0ug8I+ssksxNkiycIGvNo+IuSAcnLOPVRihvuxm3XEcGZYQxg8XWIuatoy5Ze4E87gEa6ZcwGISQ+LUhpF9CBiWE4yKMQcddjOsh/RCkg5AeSAeVxugsQfolHLeHHNpIML4dXA+jFUK+WqpujCGemwRAeD7J1CGC4U0gJen8SfxKE29ogs7+p0A6uNUmeb+D64W4QQhKkUdtnHIVt1wnW5oGlZInhffslAdfEn6ITnpFfuEqYt2K7il2797N7t27V9oMi+V1YZQi6y6RzE8Wi1+tWVTUA0B6AW5jhNLGXbjVZuHhSVkIrOMW3p7rIYwg7y2RtufJ+y3QChmUMFmC0TlOWMXxA2RYxeQZOksxUbfwlKtN3LACSNLOHDru4ZSqOGEFt9xEq5xsaYrShp24pQo6z9GyqLY8Rd5dIu8t4VWHiGeO4gQlvMY4/ZOvIL0AWaoilEH1FnGCEBGU0WmCV2kU1Z5GoaIeRimC0S0kM0fI2gtoY9B5QmnDFqTjIl0XnYgi+2IVse5F12JZzRityXvtgcieIFmYQsUdtNE4jodbGyIY3YpXH8YJK0jXQwiJ9AJwPaTjIVwP6XroLCXrzJMuTJF1FzFKI0tlpCyh8wTp+SBcpOej04isNYdwHJxShWBkG265gVFFWpnOElS/i1OuEo5swiAxQpDPn0C6AcHwhkEWz6DIRwgwBp1nJPMnkF5I1ltCpzGlrbeQLp5ESIE/tAHVXcTonLzXQQYVhDEIAW6lisqyIo488JrD4Qk6nk+yOI3OUtywhluuFL8/MOhEtXJ/wPNgRddiWUUYY1D9Lsn8JNH8JOn8SfJeqyhKcF28cpNwYhdecxS/2kQ4HsLzkZ6PdHyEW4jsqcd5AJVEJHPHi9zb7hICiQzLOLUKBo3JMsQgAVbHHfJII/0yfn0Ef2gDbqWOTvrk3UVU3C/ypb0y/lCp8IrL9SLHN4nIu4t4tSYyKJ3+nYQQRUGFMSRL05g8Q/gh6dIs4egWTJ6S9zuEG7YXYipdlE7J4jb+0EbIM4TrIrwSqrcErodbqYNRCMfBqw0TTx3CKdUINuws0iA9f3BvebqQY7VgRddiWWHyuE86f4Jo9jjpwgmy9gImTzFC4lYbhGNb8Yc34NeGkV6A8EMcNyjirF6RDXC+fGQV9UgWp0kXTpB2loqV/HINr9JEOD5Z1Eb1WugsKUQRjVcdIqgOE4xuRgZFClgyfwKdRiAkXqWBU64NHuUz3EptcG+D7rcxeYo/tPFcewSofoesNY/0A9KlGZxSDVmpkcyfwK2P4JWGSLvzyKDYr/Mcr9ZE5yleaRhpFGncxx/agNNw0UmRkuaWG7Tm/onK1lvw68OYLDnb01UXaoGzMljRtViWGZUmJPMniGeOkcwdI23No7MEIcApNwhGN+MPbSBojuGEFfACHC9Aun7h8V2i4CPvtUkWTpIsnCTvtpCej18fwWuMIPwyWWuWbHEa1VlASA/hurj1EYLmOF5jFCcok/dbxDNHT2cpePWRIvXK85Guh4l6OK6HlA5aK7RWpO15nFIVt9p4jUWyCG0sToMQ5P0uRmX4Y9vIlmZx/DL+UJENVOi3JG0tIKWDF1RJlmaKvtRag3RxymVMmhY5wSpHuB55r43wAqQXoLIE6Z7h6VKEaU6VHK80VnQtlmuMznOS+Uni2WNEM0dJFqYweYLQ4FQbBMMbCUc34Y9uwgmqOH5YeLCu92pM9BIYY8h7LZLZ4ySLJ8m7bWRYwh/ZiF8fRfphURQxdYS8u4hwHGSljl8dwmuO41WHkH6I6rWIFqfQaYL0A/zmGDIsYqSnc12NAaMR0ilUUmt0v4NOegQTO5DOGbIy+ILIWrNFzq8xqH4LvzlRZBYgCIY3n6rxRisFCPL+Eo4fgiPRaYQMKwjACQKkE2B8iZBOUaacZ4Woeh5oVTwBnLJ18P4ZoxHn7S64/FjRtViugEcfeoh333//Wdu0UiQLJ4lPHqI/c4R0YQqTJhgMXrVJaWQLwdgmgrFteKU6MggLgb2MBH5jDFlngXjqMMniFCqOcEoVSpt245brICU67hJNHyLvtECK08UQfmMMt9ooiiL6LZK5SYzKcIIK/sjGwcKcf1Z8uPj98kJ03bAQXJWTtYuwgV89t1g3j7uDijYP1WsjgyrC8dBJF39oE9IL0DofZBkUgq6jNtILIVcYo3ErTVTURgoH4QeQKIR0yVpz5FEbrzoEWqHz9LSXC5xOjUNruHr1EVeEFV2L5Qp47OGHufu++4jnThCdPEg8c5hk/iS6aJyMW6lTGttGOLaFYMNO3Ep9ECq4so+eUcXjfHTiAOnSFDpXuJU65c1bBvm4uiiU6C2i+l2M1rjNURy/hF8fLkRqIIJJbxKjFU6phlOZwA0rxWLcIEPgtWiVg5CvhhbiHnnUxa8PI8MyxpjT52mVk7bnUXmKyLOiCXm5Rh51cGvD+LVhNAaEg476SCFIey10EuMEFfIsKfJxwwrZ0gyyVEO6Pnl3CadUJZ45jjGKYGxLkUaWZ8hS9bStp0MKq6gVpxVdi+UyyOKI9kv/DMDhP/51dFr0CXDLVcLx7YTj2yht2I3XGMK5itVQWuWkC1P0Tx4kW5zGGINXHyaojRQhAAzoHK01Ku5g8hzh+LjlEm6tiV8fQUiXPOqg+h2MUTilOl59GOmHg5SzC8eMjTGgFJyaGqE1WW8JIcCvj55zblGs0cfEEUiBW26SqxQZlPGb40UYQCsEAqNzkBLVXSKLeri1sSJUUaqdjsu6YRkp3ULYpUPamcOrNAg37CA+eQCdJogzPF0AhFxVubpWdC2WN8CjDz3EYw8/fNa2P/niIwC848c/zD33ffyadDbTWUoyN0l/cj9pexYhXbzqMG6tOehn4BUetB+SpxGqvYCKivJfrzmCWxtBOi4q7pL3uyDALddxayNFDPkCGRCvxejicV86PmDIkx6q3xmkjp3dUCaPuuS9JUxWVKv5tWFwHDCGoDmO64eD7IJi+I7EkKUJaXsRHIETlnG8AJTA5BlgkH6IkKIIQSiNirr4zQlK41uJju8j6y5S3rznbKMHcefVghVdi+UN8O777z8dw027bT7/oQ/z2UceuWb3U2lMPHWY3uR+VHcJM8hL9apNZFDB8cKiGqxSFC7Ec5Oki9OAwa008GojSM8nj3vkaYRwXNz6MG51CMcLzonXXtogVTyqS4nRuiiUyFOC4Y1FVsPgMV6rogFOHnWKhS4pEa4LxuDXx4p486meooBROSpLUHEP1V/EDao4QYCREkcEqCwFRFF+bExR6ZZGIItKu2BoA0YpVNQ5p7nNasvVtaJrsVwmfrV+za6dRz36k/vpn3iFvN9CCA+3OUJQH8MpVQovtVzHLVUxWpHMTRLNHkfFnaJIoDGG8EJU2kf12+D5eM0JvNrQG8qKOJNiwnUOjoMwoNKIvNfCCcpFscKZ9ncWyTuLqDgq4qyuh5FFnq9XHyli2lKilSo6w+QZeRqRZwlGKdygjPRCpONgtMaodJAy52Ew6KRHnkS45RpykK3gBCVU3D/X8FWWq2tF12JZRaTdJbpHXyQ5cZCs30K6JcLRTXhDG3CrDbxKo2iv6BVeZdqeJzp5kGxxBuH5BCObkG5YdApLI6RfwhvbUjS8uUS89lKcCi0Ix8UIyPvtImzQGOQTU1Sf5f0ueb9N3u+gVTooQ/aRXlh43q6HGZQFCwqvWCcRGA15gpEO0ika6Qg3LJrs6KL6TDoeOo3Joy4Iid8YLRYE0wSnXCvCGWcs5BU2ra5cXSu6FssVcPd9913xNYwxpEszdA4+R3TiFXTSx602qWy9hXBkM159GLfcQPqvLhCpJCKaOkw8fbioDGuM4ATl4pE/jXBLVbz6aOEJXmGmxKs3VQOhFOg8I486COkUkyROhwmKhjzp0iwmH2Q5OBLh+vjVYbxKDWMMxhiklOg8Q+cpOs+QTiGeGIXwPKTnI07NHMScrr7Le60iz9kv4zbGEToj6yzgVppFv+B+52zPe5Xl6lrRtViugNfm6F6KM/N6jTHE00dovfw4ycxRjNL4Ixuo7rmD0vg23EoDJyid5bUZrYnnT9A/vo+8u4QMq7i1YRzHRyBwKg285ljxeH4VG3cXQqlBCIwA1W+j4wi3UsMtvbqAlnUXybqLg2bmGY4fooXAq9RxTsVxKTxik2eoPEUphXAEKIOOe6BBOAFCOgjXpfCH5cBLFWTdRQCcoIRfbRR9gHtLuNUGYt4dCPCrorvacnWt6Fosy8hjDz/Muz7yEToHn6P78rdJ2nMIx6OyaRflHbcTjGwq8mTPEwbIem26R18knZ3EGI1TquOWazhBBbfawG+M44Sla5I9UUwA1iAkmCK0AAY3rBbdyTg1or0YB1T0P/ARfoDnl/Fqw/9/e+cSI9l51fHf97j31rOnu8djexxinCAFIhJQkAEpcZAIjMSGSEhZICFPVuyQWEZiwY6FWSAiAQJFIMU2gkSRoiwiLBk2IQkWiZJITkiAJI5x/Bjb0+963Mf3fSzOreqq7pmx59HVnfT5bebRVXVrNn+dOed//kcGXAsh5nJpOOCcI7qMUO0Qy6lUtFbiKAESUTzD1hHKMc14D+MKst5Q8n7zKeX1VyQvIiuo97foPvjI/KubhUr3LKCiqygroj44AODFz3+SOB3je0PWfuHXGL77A5LMdZN+YwwN45f/l9H//Tf1wS6+2ydbvx/fv0Cxfgk/vCjh3ifIvJ9rLGEi0Y6205/bxFIINAc7VNdfJYVa/hvvvFTrWV/yemcf1tTEppbox7Y3naYj4nRCSo2Ie97BOS/PTIDzGOdlAaKucHkHP1jHGIMreoRyjLGWbLBBvXf9+D/A2DOzIKGiqygnzFFv72f/7l/mv//Q4xs89oHfuul7p1vX2P/BNymvvQTOkd/3DjqbD5JvPEA2vIjLVnOGJrULDADhYBeaRgS1HaDVox2qgy2aaiLbZJ0eWW+drDvEFH3x9gI09aF9yzqstTTlWNaJpwekkLDei4cXQ7LicrDWEcoJNI1ctih60tYwBqxvq+aGbO0io5e+RwphyTpmzNnJ1VXRVZQTZubtrScT/uKjH72pr3ex3xumE/a+/w0OXnieUJd0Lj1M76Gfo7j4EPna5kqn8CkEqUaRAV5oxJHgOn2MtZKjO96n3n6dOB3LyZ7uEN8b4ocXiaGGKC2FkCQlzbaxlCk0pNBIHm81aatpj8lzMBJqY4wFYwjlWFoWxpL1BtjugFhNpf1grYjucBNSoj7YFmfDDGPOjFf39Ed5inJOyLrdW/78K089RUqJ3R8+zyv/+hTb3/4ytuix8cu/yeYHPsLw3e+nWL9v5banFNuFCIO0FkIlCxndoVy2ONihvP4KoaowzoogDzZkCy7LRPDqStwPIKd08mL+3/0UAqGe0kxHWGvEc5t15rfcjHWkJK+RtxjJYLDtz6opNhObnF/bBKDe31r+R1grlrQzgFa6inKGeOWZf2C6fY1ssM7mr/w2g4ffi+/0TvU7pSRXH1JdyYDMiGjavKDe26LafYNmOpYIxt6QYuN+XG8N35XvnZoaUiJZcNZh80I211KSNkBKxOlYrGLGyGdbC9ZiAOOc5DfUFambcEWHrCeZvcY5Ql3iukNJPjMW1+nR7C2L7lny6qroKsoKuZmv943/fAaAan+Hjfd9mLX3PIrv3LoyXgUpRvHmxkQ1lrNBNi/w3SGhmkovd/dNwngP1xmSDTexRZ98uCEi11QiquZQcOefTWpdEYHQLkCklHBZQcLgfCZeXxJhOpKV5RjA57jubBnDijWtMyA1U8J0hB9u3rjS5Wx4dVV0FWWFHPX1HhuyfeoZQAT4Q48/fts+4HtNioEUIzEFUilJatZn2KJHtfO6tBUmo7bHK46CfP2StBRCjYmJ1AquyfN5S8G0G2mhqYhVKSltyUAC63OczzC+gCYQ64pQTyXU3EgQzlJkY4y4fo8wqgnTEdlwk/KNH8+D2OV5Z8erq6KrKKfIYoDOE1eunGh4zp0wqz6byR6RiGsHaGGyT7XzOvX4gNjUuP4Qv3aJfLCJdZ4YamwSHbXWYY7EW86CcVKKsiBRTUgE6fdmGSDDtsYiN9xCg0lytNL3DhcfYlNhvMNYj827hHIswzSkr1tcvAycLa+uDtIURbkhs9ZCjIEw2pehlsswPqfafZN6/zqpHGGLLvnwEnn/gvhxY4NJEEnYeUtBEsXSoVtX+sQxEMsRsa6wOPHj+hy8E2dENWntXxm0Q7r50UlEkA1WFiryjrQ/ih4Yc7zFcEZsYyq6inJGuBc5DveS2QAtTGWpw7QVa5geUG6/JgcmcfjeBfxwHd9fByJEcTrMBHfm711aTmjzF0iJZjxq83Jbi5jPsV6GbbGayPn5lLAul6B2e9gfSHWFyYu2QhYxTk2F7184tiRhzkiYuYquopwRTrt/e4wQIETCaA+cFYsWifL6KxKEHhp8d4i/cBE/2CCZBDFhbGv1yhYuOMyyc2fCm9Lcn0soW1tajc0KnPUS4dhUEOL8xLvx/pjoxrqUEHbrxPtr3byvW+9vzdsYQGsbO/2tNBVdRVGOMQu4qatxm4ebE2OgHu1S720R6ykm6+AubJL317FZhsEsCe5SBsSC2ELrXCDRTMeEairJZTipkIuOJI6VI3Eb+IwUZQvNGOaim0IgxSA5u87Ja7POXHRTUxMmB/OvYMzZ8Oqq6CqKcoyZayGM9to120gox1Q7b8oFimTxg3WKwQauNxDBNdxYcFvM0SKzaUjTMaGpwHsMEeM9pn0eob2dluTvfSG+35noxlqOf9qskOcai80KObI5W09e7OvOwnZOua+roqsoynFCINQVsRpj85wwGVHvbRPKsfRXO12ywTpusNEK3rLgHhPdo39OSU6mlxJ0Y4zBuBznC0yWEcrxPDBdWhb20OPbZirEppKPzmZn4q0M8lKSAHTnl0X3jDgY1DKm3s/5vgAAC75JREFUKMoSqXUsNAc7ckk3RKqda6RQQwKsrPlmF+5vsxBuXeEeFdwY2wFdDDTTAyCQjMO1zgWTkPaF9Ug3N+Hyw0WRxUrX+oVnOoeN0tuN5Rg/3FjaTDsrXl2tdBVFWSZGUl0TpgcYl1NtvUwY72GMI4aALTp0Nh7A9YfYtxLcG9FaxVKoSU1Diql1PFhs0SM0U1KM89wGkxK225fXsNDTraulYZ21jmTAZMXcr1uPdkht5sNZ8eqq6CqKskSKgXqyi8UQqzHT11+SQJkYwUBn8zLZhUtY0u0JrlmwjoVAmI5EeK2ZC6rPChFTmZhJnkJ7Y02GbwZjLTE08yHaHGvlekbeOezrxkgz2l34Dqfv1VXRVRRliVBXhPEB0TomL39fqs5On1BP6Wxcprj0MMa8PcFNRyxaM29uDDXNaI8YA8a0rgWXyYaZy0it/SulKIMyjAjwQpULLC1KSF/YY7KO3GDLJNh90a97Fry6KrqKosxJIcgpnpQot16l3L5Gtv4A1cEuvtOnePBdOO9ktfd2WgpweLwyyemfODkgNbUcnbSO1IquCLCEl6eUsHl7FSOlwyFa61xY8gIjqWPGORHnFLFF9/gwTStdRVHOCjE0hPEeoRwzeeUH2O6g7as29B56D77bu33BnS1GtCQSsS4JVYlJUayz1mCMiKpJUsGmEFo72Eys0+EQramWh2jzR0nguc2LpSWJpZ9zugsSKrqKosxpxvs05YTxyz8glGP88D7iaJvOpUfINi5h3e1XuEdJTSP5uM2UZJBerHFS9fpc/uwzSBFX9GXQ1m6zHbYXymNVLsiwzBrX9qCDtEUmB8S2HTG3jZ1itauiqygKIELUjHaYXvsR5d51io0HSNUY21uje/ld+Cxbqjzfihu9LrXrv2G8RwiNVKWzlV+XYY20BgytgBZdQHIaZttoMkSLS/3cJZyT96U09/bOq92FFsdpoaKrKAqAVLjXXmT65qv4oktIIsT9d76XrD+4LcG9FSk0NNORbJvFIBeGq1K8uNa2geQVNu+JT3dBII11h0M0f7zShTZKsh3EWSeRkjPRXfLqnhIquoqiAPDvf/8pxj/6L8lU6K8Tp2O6DzxM99JD90xwSYlYV8TJCGLEptn1iITrdKWa9TkpRlx3AO11iXl2g3MyRDPmhu0FYG4ds3khVyV6Q5qZ6J4Br66KrqIoxKbha1/4IjEGvvf8i9TXf0wxXKf/yPvuXnAX1oJNglCOiPUEgGQsMUxxPse6rLV5RazPWr9tnH8GtO2FpsK67KbfyRgjVW7WJcaA6w6oFmMeT9mrq6KrKArb3/saAN2f/UW+9W9fJRnL8Od/lbw3uOsKd9GrG2JDM96TvqwBvCeVlRyWjBHX6RPLyTywRt4fIc62ypwM0fKb9HNbjPeYQkTb5h05/z4Zyc9O2aur2QuKco45eqPt03/y5wB85m++wAf3hnz44x+/Z8+aXYpoDvZJbb/YxkCIDS7LMD6ThDGqtrUwf2PbXjBzj+/N+rkzjLFYn2Gcw5rZMO26HLS0Fpr6nv27bhetdBXlHPPY1as3vVjx1aef5stPPnnvHmYMhIZmvAvWkJqGpinBF+Acrn+BFCTI3LY31RKzCxNtlVsdxjne8lHWYrHSrrCyaFHvLQ7T0rFtuVWhoqso55zHrl6dH8T848/8EwCfePZZPvHss/f0moUcuGz7uSnJtWAj223W57hOTy5BLFW5QIqSu+Bcu7Fmjh26vCHeL+UwzG1ji5eETwEVXUVR5nQ277vnn7l4I60e75BCxBorWb1ZhxSmuO5AXmUMrg0rPzzvk+aLEcfiHG+BtQ5b9GSzLe9QH2yTYpzbxk6rr6uiqygKcDKHMRfFMaVEHO1BCjRNjbFZmyRm2/M61dyru0RKpFk+Q1Pd3Cp2FNv2ddvbavPEsdl3OiUHg4quoijA4WHMk7pKnEJDfbCNyQviaAdXdEjNFJt35qE2S60FZNiWYnu2J8Zbb6IdYW4dyzvi/SVR72+duldXRVdRlCVO6ipxU05oxvuAI5RTWcAoJ/jBGpAO136XTrW3ixHGzG1jb+VcWMI5XN6VixSYw5jHU/TqqugqirIS4mgHQg2xlm20LCeGGjfcFNdCpz+vQufviYEYAsYaUpvV8LaGaC3WeWynD7FZink8Ta+uiq6iKCeLMaSUqHbfJDSNDNLyjtjAjCXLexjj8J3B8ffGIOLYiuTbHaIdPtrgXIZxGTbvEsb7xKY+1VxdFV1FUU6cGKOs4hpHqqeYrCDVJfhMBl55gc2Xs3NBfLokuVKRQvP2h2iLOIcruhifHfZ1T9Grq6KrKMqJk5qGZn8b4z0xJlxREOoS4zoY55bWfpfeF4NcmWir0rc7RFvEOo/Nu2Ifq0q5EGxPL21MRVdRlBOnPtiiHu/iOwNiKMHnElLuDDYrll0LM3+uMa0oJhmocZtDtNnHOYftDrDWYny+UOm2lfSKUdFVFOVEMcZQbb3Ot5/7DrbbIzUlrrtOjA3Ge3x3KDfRboBkLQQk1fz2hmiLOJ9jshybFzJM00pXUZSfWoxh8saL/Nc3f4h1juef+y7Z2iapmmDzHr6/dsO3pZTkNHtKENJtD9GWcA6XdTC+IJRjYtlGS56Cg0FFV1GUE2f65ssANHu7fPebL2CLnGZyQNa/cLj227IkrClJtUu6o37ujJl1zHX6pLqSaveUvLoa7agoyolwNDYS4Mk/+1sAvvb5LzJ97QU+8uu/e8ybu0gKtbgWrL0z50KLMUbaGEWH2NTUe9fJBhta6SqK8tPDLL3sE88+yx/98z8C8LGP/wYAv/TB9/Pdb71AcfHyjd9sJCYnNTXE0F6CuPNKF8D4DFd0l/u62tNVFOWnjS8/+SR/9ft/AMDnPv0lAD79p58E4LnPfu6W742hJsUG6/08Y/dOsc7jip44GPauI9O51Xt1VXQVRTlRFvN6j/LVp5/miStXjoWlz/q6IdTtBlv3rr+Hca7t6/YI5YRmsi8/WHG1q6KrKMpK+NDjj8/Fd/HXm4WlGyPXJVK6N6IL4DtDXHeNVJeEgx1g9Q4GFV1FUVbCnURHxmqCsQ7XRj/eLSbPyYYbxNBQj0R0V31BQkVXUZSV8rajI40hViXPP/edu3IuLCJ93S52oa+rla6iKOeCt1PxpnLC8//x7bseos0wxuB6Q1x/SLXzBjFF7ekqinI+eKuK1xhDaMp7/lzXXcN3B8RqStjfXnmlq8sRiqKcKW60VPHElSuAVMd3e9nCOk+2/iDjl/6HerRLtnbxrj7vdlHRVRTlTPHY1atzYT149Uf89dU/vKnl7E6w3pMN1sE6iXm8/C5SSnee63C7z1/JUxRFUe6AweVHTuRzfXcN1+lT7lyTv1hhX1dFV1GUc4ftDfD9NerRLqGcrLSvq6KrKMqZ5iROwrssJ9+4TKomNAfbWukqiqLMOImT8MYY8s0HiKGm2t3SSldRFOWkyQab+KJPuf3aSrfSVHQVRTmXuO4A179AtfUaMYaVPVdFV1GUc4nLMvL1B4jTg3n4zSpQ0VUU5dxSXPoZYlNT7b65slxdFV1FUc4t+eZljPNyw21FDgYVXUVRzi2+N8QP16m2XlmZg0FFV1GUc4tzjnzzMuXOG8S6WskzVXQVRTnXdO97JzQV1e4bK3meiq6iKOea4v6HwRi+8uRTb/3ie4CKrqIo55r8wkVsp883nv3SSp6noqsoyrnGGEOx+eDKnqeiqyjKuadz8R0AlAe7J/4sDTFXFOXccvRKxV/+3sfmv78XVypuhHmLLYzV3iZWlHvEo48+yte//vXT/hrKTxBPXLlyLy9U3PQMhbYXFEVRVoiKrqIoCicTln4jVHQVRVE4mbD0G6GiqyiKskJUdBVFUVaIiq6iKMoKUdFVFEVZISq6iqIoK0RFV1EUZYWo6CqKoqwQFV1FUZQV8lbZC4ryE4kx5pmU0u+c9vdQlKOo6CqKoqwQbS8oiqKsEBVdRVGUFaKiqyiKskJUdBVFUVaIiq6iKMoK+X9qqOqp4QPDwwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "toy_dataset.main(algoName=\"MLOT\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Cheat tables"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This can be used to recreate all the table in the paper. However to allow more complex cross validation, another code is provided above. It will launch again all the method with the best hyperparameters set, this is essential for random methods.\n",
    "The code bellow can be run only if the cross validation file are already saved."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import recreate_table"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\\newpage Feature :$ surf_SA $ Cheat  True \\newline\n",
      "\\begin{tabular}{lrrrrrlll}\n",
      "\\toprule\n",
      "         dataset &  JDOTSVMe &      OT &    TCA &    LMNN &     SA &           OTDA &       OTDA_pca &           MLOT \\\\\n",
      "\\midrule\n",
      " A$\\rightarrow$C &      41.5 &    33.8 &   40.0 &    41.9 &   40.5 &           42.2 &           42.6 &  \\textbf{43.5} \\\\\n",
      " A$\\rightarrow$D &      42.0 &    34.4 &   33.1 &    42.7 &   40.6 &           47.1 &           49.2 &  \\textbf{50.6} \\\\\n",
      " A$\\rightarrow$W &      38.6 &    33.9 &   41.7 &    40.7 &   41.6 &           46.4 &           50.2 &  \\textbf{51.3} \\\\\n",
      " C$\\rightarrow$A &      48.9 &    45.2 &   40.6 &    48.4 &   42.3 &           53.3 &           55.8 &  \\textbf{57.8} \\\\\n",
      " C$\\rightarrow$D &      49.7 &    44.6 &   44.6 &    47.1 &   46.8 &           55.4 &           56.5 &  \\textbf{57.3} \\\\\n",
      " C$\\rightarrow$W &      43.4 &    38.6 &   38.6 &    44.4 &   41.0 &           53.9 &           55.3 &  \\textbf{59.2} \\\\\n",
      " D$\\rightarrow$A &      33.6 &    30.7 &   34.0 &    37.9 &   36.4 &           38.4 &           40.8 &  \\textbf{42.5} \\\\\n",
      " D$\\rightarrow$C &      33.6 &    32.3 &   32.2 &    33.9 &   33.1 &  \\textbf{36.5} &           36.2 &           36.4 \\\\\n",
      " D$\\rightarrow$W &      86.8 &    88.8 &   86.8 &    77.3 &   89.8 &  \\textbf{91.2} &           90.3 &           90.4 \\\\\n",
      " W$\\rightarrow$A &      38.4 &    39.0 &   30.1 &    36.4 &   34.6 &           41.2 &           41.8 &  \\textbf{42.8} \\\\\n",
      " W$\\rightarrow$C &      34.9 &    32.6 &   30.0 &    32.9 &   31.8 &           35.7 &  \\textbf{36.5} &           36.2 \\\\\n",
      " W$\\rightarrow$D &      90.4 &    89.8 &   92.4 &    91.1 &   90.3 &           92.4 &           92.9 &  \\textbf{93.0} \\\\\n",
      "             AVG &      48.5 &    45.3 &   45.3 &    47.9 &   47.4 &           52.8 &             54 &  \\textbf{55.1} \\\\\n",
      "              nb &    1000.0 &  1000.0 &  884.7 &  1000.0 &  899.4 &           1000 &           1000 &            628 \\\\\n",
      "\\bottomrule\n",
      "\\end{tabular}\n",
      "\n",
      "\\newpage Feature :$ surf_SA $ Cheat  False \\newline\n",
      "\\begin{tabular}{llrlrllrl}\n",
      "\\toprule\n",
      "         dataset &       JDOTSVMe &      OT &            TCA &    LMNN &             SA &           OTDA &  OTDA_pca &           MLOT \\\\\n",
      "\\midrule\n",
      " A$\\rightarrow$C &             40 &    33.8 &             40 &    40.3 &           40.1 &           40.2 &      38.9 &  \\textbf{42.1} \\\\\n",
      " A$\\rightarrow$D &           37.6 &    30.6 &           31.8 &    36.9 &           39.6 &           40.1 &      39.1 &  \\textbf{40.3} \\\\\n",
      " A$\\rightarrow$W &             38 &    32.5 &  \\textbf{41.7} &    38.0 &           39.8 &           37.3 &      40.5 &           41.5 \\\\\n",
      " C$\\rightarrow$A &           48.4 &    41.0 &           39.8 &    46.0 &           41.1 &  \\textbf{52.7} &      48.6 &           51.6 \\\\\n",
      " C$\\rightarrow$D &           49.7 &    36.9 &           44.6 &    45.9 &           46.4 &           47.8 &      50.7 &  \\textbf{52.7} \\\\\n",
      " C$\\rightarrow$W &           43.4 &    28.1 &           38.3 &    41.7 &           37.4 &  \\textbf{46.4} &      45.3 &           45.7 \\\\\n",
      " D$\\rightarrow$A &           32.4 &    29.3 &           32.9 &    31.1 &           34.7 &           32.4 &      38.0 &  \\textbf{38.0} \\\\\n",
      " D$\\rightarrow$C &           31.9 &    31.7 &           31.5 &    30.7 &           32.6 &             32 &      33.7 &  \\textbf{34.4} \\\\\n",
      " D$\\rightarrow$W &           82.7 &    88.8 &           84.7 &    77.3 &  \\textbf{89.2} &           88.8 &      88.0 &             88 \\\\\n",
      " W$\\rightarrow$A &  \\textbf{37.7} &    34.1 &           29.4 &    32.3 &           32.5 &           33.2 &      37.0 &           37.6 \\\\\n",
      " W$\\rightarrow$C &           33.2 &    30.1 &           29.2 &    30.4 &           28.9 &  \\textbf{34.1} &      33.5 &           33.1 \\\\\n",
      " W$\\rightarrow$D &           89.8 &    89.2 &           91.7 &    86.6 &           89.3 &  \\textbf{92.4} &      92.1 &           90.1 \\\\\n",
      "             AVG &           47.1 &    42.2 &           44.6 &    44.8 &             46 &           48.1 &      48.8 &  \\textbf{49.6} \\\\\n",
      "              nb &           1000 &  1000.0 &          884.7 &  1000.0 &          899.4 &           1000 &    1000.0 &            628 \\\\\n",
      "\\bottomrule\n",
      "\\end{tabular}\n",
      "\n",
      "\\newpage Feature :$ decaf6_SA $ Cheat  True \\newline\n",
      "\\begin{tabular}{llrlrrrll}\n",
      "\\toprule\n",
      "         dataset &        JDOTSVMe &      OT &            TCA &    LMNN &     SA &    OTDA &       OTDA_pca &           MLOT \\\\\n",
      "\\midrule\n",
      " A$\\rightarrow$C &            85.6 &    84.6 &           78.5 &    83.1 &   81.5 &    88.0 &  \\textbf{88.1} &           87.5 \\\\\n",
      " A$\\rightarrow$D &            87.9 &    79.6 &           84.1 &    85.4 &   83.9 &    91.1 &  \\textbf{92.4} &           90.7 \\\\\n",
      " A$\\rightarrow$W &            82.7 &    76.3 &           73.2 &    76.6 &   75.3 &    89.2 &  \\textbf{91.1} &           89.4 \\\\\n",
      " C$\\rightarrow$A &            91.5 &    90.4 &           89.6 &    92.3 &   88.3 &    93.1 &  \\textbf{93.5} &           92.9 \\\\\n",
      " C$\\rightarrow$D &            88.5 &    89.8 &           82.8 &    89.8 &   87.5 &    90.4 &  \\textbf{92.1} &           91.3 \\\\\n",
      " C$\\rightarrow$W &            78.3 &    81.4 &           81.4 &    80.3 &   86.0 &    88.8 &  \\textbf{92.8} &           89.8 \\\\\n",
      " D$\\rightarrow$A &            86.8 &    89.9 &           86.7 &    73.6 &   84.8 &    92.5 &  \\textbf{92.9} &           92.7 \\\\\n",
      " D$\\rightarrow$C &            81.6 &    85.1 &           77.7 &    70.2 &   78.3 &    87.5 &  \\textbf{88.2} &           88.1 \\\\\n",
      " D$\\rightarrow$W &              99 &    99.3 &  \\textbf{99.7} &    99.3 &   99.5 &    99.3 &           99.3 &           98.1 \\\\\n",
      " W$\\rightarrow$A &            86.2 &    88.2 &           81.8 &    74.1 &   84.9 &    91.2 &           92.4 &  \\textbf{92.8} \\\\\n",
      " W$\\rightarrow$C &            79.6 &    82.6 &             74 &    67.5 &   76.4 &    86.7 &           87.4 &  \\textbf{87.6} \\\\\n",
      " W$\\rightarrow$D &  \\textbf{100.0} &   100.0 &            100 &   100.0 &  100.0 &   100.0 &            100 &           99.2 \\\\\n",
      "             AVG &            87.3 &    87.3 &           84.1 &    82.7 &   85.5 &    91.5 &  \\textbf{92.5} &           91.7 \\\\\n",
      "              nb &            1000 &  1000.0 &            911 &  1000.0 &  724.4 &  1000.0 &           1000 &           22.8 \\\\\n",
      "\\bottomrule\n",
      "\\end{tabular}\n",
      "\n",
      "\\newpage Feature :$ decaf6_SA $ Cheat  False \\newline\n",
      "\\begin{tabular}{llllllrrl}\n",
      "\\toprule\n",
      "         dataset &        JDOTSVMe &             OT &            TCA &           LMNN &             SA &    OTDA &  OTDA_pca &           MLOT \\\\\n",
      "\\midrule\n",
      " A$\\rightarrow$C &   \\textbf{83.0} &           82.2 &           77.7 &           80.6 &           81.1 &    80.5 &      79.5 &           81.5 \\\\\n",
      " A$\\rightarrow$D &            70.7 &           77.7 &  \\textbf{84.1} &           80.3 &           83.4 &    76.4 &      74.1 &           74.6 \\\\\n",
      " A$\\rightarrow$W &            76.9 &           71.2 &           71.9 &           72.5 &           74.8 &    71.9 &      74.9 &  \\textbf{77.2} \\\\\n",
      " C$\\rightarrow$A &            90.5 &           90.4 &           89.1 &  \\textbf{91.4} &           87.8 &    86.8 &      87.4 &           90.8 \\\\\n",
      " C$\\rightarrow$D &            85.4 &           82.8 &           82.8 &             79 &  \\textbf{86.9} &    80.9 &      80.5 &           83.1 \\\\\n",
      " C$\\rightarrow$W &            76.9 &             78 &           79.7 &             78 &  \\textbf{85.8} &    80.3 &      77.8 &           77.7 \\\\\n",
      " D$\\rightarrow$A &            64.9 &           84.4 &             86 &           70.1 &           84.3 &    84.2 &      82.7 &  \\textbf{87.3} \\\\\n",
      " D$\\rightarrow$C &            64.7 &           81.2 &           75.2 &           63.7 &           72.4 &    80.9 &      81.1 &  \\textbf{83.4} \\\\\n",
      " D$\\rightarrow$W &            98.6 &  \\textbf{99.3} &             99 &           95.6 &           99.2 &    98.3 &      98.6 &           96.5 \\\\\n",
      " W$\\rightarrow$A &            64.3 &           81.5 &           81.8 &           73.4 &             76 &    81.3 &      77.4 &  \\textbf{87.8} \\\\\n",
      " W$\\rightarrow$C &            61.9 &           78.6 &           73.6 &             67 &           72.9 &    75.7 &      77.6 &  \\textbf{81.0} \\\\\n",
      " W$\\rightarrow$D &  \\textbf{100.0} &            100 &            100 &            100 &            100 &   100.0 &      99.2 &           99.2 \\\\\n",
      "             AVG &            78.2 &           83.9 &           83.4 &           79.3 &           83.7 &    83.1 &      82.6 &  \\textbf{85.0} \\\\\n",
      "              nb &            1000 &           1000 &            911 &           1000 &          724.4 &  1000.0 &    1000.0 &           22.8 \\\\\n",
      "\\bottomrule\n",
      "\\end{tabular}\n",
      "\n",
      "\\newpage Feature :$ office31fc7_SA $ Cheat  True \\newline\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\\begin{tabular}{lrrrlrlll}\n",
      "\\toprule\n",
      "         dataset &  JDOTSVMe &      OT &     TCA &           LMNN &      SA &           OTDA &       OTDA_pca &           MLOT \\\\\n",
      "\\midrule\n",
      " A$\\rightarrow$D &      59.2 &    56.8 &    58.4 &           60.6 &    61.4 &  \\textbf{64.9} &           64.5 &           63.3 \\\\\n",
      " A$\\rightarrow$W &      60.4 &    54.8 &    52.7 &           55.7 &    56.6 &           66.4 &  \\textbf{67.5} &           65.3 \\\\\n",
      " D$\\rightarrow$A &      50.8 &    51.1 &    44.1 &           48.3 &    45.5 &           54.3 &           54.4 &  \\textbf{57.9} \\\\\n",
      " D$\\rightarrow$W &      91.7 &    94.7 &    91.7 &           95.6 &    93.3 &             95 &             95 &  \\textbf{95.7} \\\\\n",
      " W$\\rightarrow$A &      49.4 &    49.7 &    42.7 &           45.9 &    45.5 &           52.7 &           52.7 &  \\textbf{56.2} \\\\\n",
      " W$\\rightarrow$D &      95.0 &    97.2 &    95.8 &  \\textbf{99.2} &    98.8 &           97.2 &           97.2 &           97.8 \\\\\n",
      "             AVG &      67.8 &    67.4 &    64.2 &           67.6 &    66.9 &           71.7 &           71.9 &  \\textbf{72.7} \\\\\n",
      "              nb &     767.8 &  1000.0 &  1000.0 &           1000 &  1000.0 &           1000 &           1000 &            286 \\\\\n",
      "\\bottomrule\n",
      "\\end{tabular}\n",
      "\n",
      "\\newpage Feature :$ office31fc7_SA $ Cheat  False \\newline\n",
      "\\begin{tabular}{lrlrlllrl}\n",
      "\\toprule\n",
      "         dataset &  JDOTSVMe &             OT &     TCA &           LMNN &             SA &           OTDA &  OTDA_pca &           MLOT \\\\\n",
      "\\midrule\n",
      " A$\\rightarrow$D &      57.8 &           56.8 &    58.0 &           50.6 &  \\textbf{61.0} &           55.4 &      53.6 &           54.8 \\\\\n",
      " A$\\rightarrow$W &      52.1 &           49.2 &    52.7 &           51.8 &  \\textbf{56.4} &           50.8 &      53.2 &           53.7 \\\\\n",
      " D$\\rightarrow$A &      45.8 &           48.8 &    44.0 &             48 &             45 &           48.1 &      47.4 &  \\textbf{50.4} \\\\\n",
      " D$\\rightarrow$W &      90.9 &           92.1 &    91.7 &           94.7 &           93.2 &  \\textbf{95.0} &      95.0 &           92.6 \\\\\n",
      " W$\\rightarrow$A &      44.8 &  \\textbf{48.3} &    42.5 &           43.8 &           45.1 &           46.9 &      46.2 &           47.6 \\\\\n",
      " W$\\rightarrow$D &      94.8 &           96.8 &    95.6 &  \\textbf{99.0} &           98.6 &           95.4 &      96.0 &           97.8 \\\\\n",
      "             AVG &      64.4 &           65.3 &    64.1 &           64.7 &  \\textbf{66.5} &           65.3 &      65.2 &           66.2 \\\\\n",
      "              nb &     767.8 &           1000 &  1000.0 &           1000 &           1000 &           1000 &    1000.0 &            286 \\\\\n",
      "\\bottomrule\n",
      "\\end{tabular}\n",
      "\n"
     ]
    }
   ],
   "source": [
    "recreate_table.creat_all_cheat_tables(\"JDOTSVMe,OT,TCA,LMNN,SA,OTDA,OTDA_pca,MLOT\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\\newpage Feature :$ surf_SA $ Cheat  True \\newline\n",
      "\\begin{tabular}{lrrrrrll}\n",
      "\\toprule\n",
      "         dataset &    NA &     SA &  CORAL &    TCA &      OT &           OTDA &           MLOT \\\\\n",
      "\\midrule\n",
      " A$\\rightarrow$C &  26.0 &   40.5 &   25.4 &   40.0 &    33.8 &           42.2 &  \\textbf{43.5} \\\\\n",
      " A$\\rightarrow$D &  25.5 &   40.6 &   26.8 &   33.1 &    34.4 &           47.1 &  \\textbf{50.6} \\\\\n",
      " A$\\rightarrow$W &  29.8 &   41.6 &   26.8 &   41.7 &    33.9 &           46.4 &  \\textbf{51.3} \\\\\n",
      " C$\\rightarrow$A &  23.7 &   42.3 &   23.6 &   40.6 &    45.2 &           53.3 &  \\textbf{57.8} \\\\\n",
      " C$\\rightarrow$D &  25.5 &   46.8 &   26.1 &   44.6 &    44.6 &           55.4 &  \\textbf{57.3} \\\\\n",
      " C$\\rightarrow$W &  25.8 &   41.0 &   23.7 &   38.6 &    38.6 &           53.9 &  \\textbf{59.2} \\\\\n",
      " D$\\rightarrow$A &  28.5 &   36.4 &   28.8 &   34.0 &    30.7 &           38.4 &  \\textbf{42.5} \\\\\n",
      " D$\\rightarrow$C &  26.3 &   33.1 &   30.0 &   32.2 &    32.3 &  \\textbf{36.5} &           36.4 \\\\\n",
      " D$\\rightarrow$W &  63.4 &   89.8 &   84.4 &   86.8 &    88.8 &  \\textbf{91.2} &           90.4 \\\\\n",
      " W$\\rightarrow$A &  23.0 &   34.6 &   26.2 &   30.1 &    39.0 &           41.2 &  \\textbf{42.8} \\\\\n",
      " W$\\rightarrow$C &  19.9 &   31.8 &   22.6 &   30.0 &    32.6 &           35.7 &  \\textbf{36.2} \\\\\n",
      " W$\\rightarrow$D &  59.2 &   90.3 &   84.1 &   92.4 &    89.8 &           92.4 &  \\textbf{93.0} \\\\\n",
      "             AVG &  31.4 &   47.4 &   35.7 &   45.3 &    45.3 &           52.8 &  \\textbf{55.1} \\\\\n",
      "              nb &   2.0 &  899.4 &  986.8 &  884.7 &  1000.0 &           1000 &            628 \\\\\n",
      "\\bottomrule\n",
      "\\end{tabular}\n",
      "\n",
      "\\newpage Feature :$ surf_SA $ Cheat  False \\newline\n",
      "\\begin{tabular}{lrlrlrll}\n",
      "\\toprule\n",
      "         dataset &    NA &             SA &  CORAL &            TCA &      OT &           OTDA &           MLOT \\\\\n",
      "\\midrule\n",
      " A$\\rightarrow$C &  26.0 &           40.1 &   25.4 &             40 &    33.8 &           40.2 &  \\textbf{42.1} \\\\\n",
      " A$\\rightarrow$D &  25.5 &           39.6 &   26.8 &           31.8 &    30.6 &           40.1 &  \\textbf{40.3} \\\\\n",
      " A$\\rightarrow$W &  29.8 &           39.8 &   26.8 &  \\textbf{41.7} &    32.5 &           37.3 &           41.5 \\\\\n",
      " C$\\rightarrow$A &  23.7 &           41.1 &   23.6 &           39.8 &    41.0 &  \\textbf{52.7} &           51.6 \\\\\n",
      " C$\\rightarrow$D &  25.5 &           46.4 &   26.1 &           44.6 &    36.9 &           47.8 &  \\textbf{52.7} \\\\\n",
      " C$\\rightarrow$W &  25.8 &           37.4 &   23.7 &           38.3 &    28.1 &  \\textbf{46.4} &           45.7 \\\\\n",
      " D$\\rightarrow$A &  28.5 &           34.7 &   28.8 &           32.9 &    29.3 &           32.4 &  \\textbf{38.0} \\\\\n",
      " D$\\rightarrow$C &  26.3 &           32.6 &   30.0 &           31.5 &    31.7 &             32 &  \\textbf{34.4} \\\\\n",
      " D$\\rightarrow$W &  63.4 &  \\textbf{89.2} &   84.4 &           84.7 &    88.8 &           88.8 &             88 \\\\\n",
      " W$\\rightarrow$A &  23.0 &           32.5 &   26.2 &           29.4 &    34.1 &           33.2 &  \\textbf{37.6} \\\\\n",
      " W$\\rightarrow$C &  19.9 &           28.9 &   22.6 &           29.2 &    30.1 &  \\textbf{34.1} &           33.1 \\\\\n",
      " W$\\rightarrow$D &  59.2 &           89.3 &   84.1 &           91.7 &    89.2 &  \\textbf{92.4} &           90.1 \\\\\n",
      "             AVG &  31.4 &             46 &   35.7 &           44.6 &    42.2 &           48.1 &  \\textbf{49.6} \\\\\n",
      "              nb &   2.0 &          899.4 &  986.8 &          884.7 &  1000.0 &           1000 &            628 \\\\\n",
      "\\bottomrule\n",
      "\\end{tabular}\n",
      "\n",
      "\\newpage Feature :$ decaf6_SA $ Cheat  True \\newline\n",
      "\\begin{tabular}{lrlrlrll}\n",
      "\\toprule\n",
      "         dataset &    NA &              SA &  CORAL &            TCA &      OT &           OTDA &           MLOT \\\\\n",
      "\\midrule\n",
      " A$\\rightarrow$C &  71.7 &            81.5 &   72.2 &           78.5 &    84.6 &  \\textbf{88.0} &           87.5 \\\\\n",
      " A$\\rightarrow$D &  73.9 &            83.9 &   72.0 &           84.1 &    79.6 &  \\textbf{91.1} &           90.7 \\\\\n",
      " A$\\rightarrow$W &  68.1 &            75.3 &   64.1 &           73.2 &    76.3 &           89.2 &  \\textbf{89.4} \\\\\n",
      " C$\\rightarrow$A &  87.3 &            88.3 &   81.3 &           89.6 &    90.4 &  \\textbf{93.1} &           92.9 \\\\\n",
      " C$\\rightarrow$D &  79.6 &            87.5 &   80.3 &           82.8 &    89.8 &           90.4 &  \\textbf{91.3} \\\\\n",
      " C$\\rightarrow$W &  72.5 &              86 &   68.8 &           81.4 &    81.4 &           88.8 &  \\textbf{89.8} \\\\\n",
      " D$\\rightarrow$A &  49.9 &            84.8 &   78.0 &           86.7 &    89.9 &           92.5 &  \\textbf{92.7} \\\\\n",
      " D$\\rightarrow$C &  42.0 &            78.3 &   71.8 &           77.7 &    85.1 &           87.5 &  \\textbf{88.1} \\\\\n",
      " D$\\rightarrow$W &  91.5 &            99.5 &   99.3 &  \\textbf{99.7} &    99.3 &           99.3 &           98.1 \\\\\n",
      " W$\\rightarrow$A &  62.5 &            84.9 &   69.7 &           81.8 &    88.2 &           91.2 &  \\textbf{92.8} \\\\\n",
      " W$\\rightarrow$C &  55.3 &            76.4 &   69.1 &             74 &    82.6 &           86.7 &  \\textbf{87.6} \\\\\n",
      " W$\\rightarrow$D &  98.1 &  \\textbf{100.0} &  100.0 &            100 &   100.0 &            100 &           99.2 \\\\\n",
      "             AVG &  71.0 &            85.5 &   77.2 &           84.1 &    87.3 &           91.5 &  \\textbf{91.7} \\\\\n",
      "              nb &   2.3 &           724.4 &  308.7 &            911 &  1000.0 &           1000 &           22.8 \\\\\n",
      "\\bottomrule\n",
      "\\end{tabular}\n",
      "\n",
      "\\newpage Feature :$ decaf6_SA $ Cheat  False \\newline\n",
      "\\begin{tabular}{lrllllrl}\n",
      "\\toprule\n",
      "         dataset &    NA &              SA &          CORAL &            TCA &             OT &    OTDA &           MLOT \\\\\n",
      "\\midrule\n",
      " A$\\rightarrow$C &  71.7 &            81.1 &           72.2 &           77.7 &  \\textbf{82.2} &    80.5 &           81.5 \\\\\n",
      " A$\\rightarrow$D &  73.9 &            83.4 &             72 &  \\textbf{84.1} &           77.7 &    76.4 &           74.6 \\\\\n",
      " A$\\rightarrow$W &  68.1 &            74.8 &           64.1 &           71.9 &           71.2 &    71.9 &  \\textbf{77.2} \\\\\n",
      " C$\\rightarrow$A &  87.3 &            87.8 &           81.3 &           89.1 &           90.4 &    86.8 &  \\textbf{90.8} \\\\\n",
      " C$\\rightarrow$D &  79.6 &   \\textbf{86.9} &           80.3 &           82.8 &           82.8 &    80.9 &           83.1 \\\\\n",
      " C$\\rightarrow$W &  72.5 &   \\textbf{85.8} &           68.8 &           79.7 &             78 &    80.3 &           77.7 \\\\\n",
      " D$\\rightarrow$A &  49.9 &            84.3 &             78 &             86 &           84.4 &    84.2 &  \\textbf{87.3} \\\\\n",
      " D$\\rightarrow$C &  42.0 &            72.4 &           71.8 &           75.2 &           81.2 &    80.9 &  \\textbf{83.4} \\\\\n",
      " D$\\rightarrow$W &  91.5 &            99.2 &  \\textbf{99.3} &             99 &           99.3 &    98.3 &           96.5 \\\\\n",
      " W$\\rightarrow$A &  62.5 &              76 &           69.7 &           81.8 &           81.5 &    81.3 &  \\textbf{87.8} \\\\\n",
      " W$\\rightarrow$C &  55.3 &            72.9 &           69.1 &           73.6 &           78.6 &    75.7 &  \\textbf{81.0} \\\\\n",
      " W$\\rightarrow$D &  98.1 &  \\textbf{100.0} &            100 &            100 &            100 &   100.0 &           99.2 \\\\\n",
      "             AVG &  71.0 &            83.7 &           77.2 &           83.4 &           83.9 &    83.1 &  \\textbf{85.0} \\\\\n",
      "              nb &   2.3 &           724.4 &          308.7 &            911 &           1000 &  1000.0 &           22.8 \\\\\n",
      "\\bottomrule\n",
      "\\end{tabular}\n",
      "\n",
      "\\newpage Feature :$ office31fc7_SA $ Cheat  True \\newline\n",
      "\\begin{tabular}{lrlrrrll}\n",
      "\\toprule\n",
      "         dataset &    NA &             SA &  CORAL &     TCA &      OT &           OTDA &           MLOT \\\\\n",
      "\\midrule\n",
      " A$\\rightarrow$D &  59.6 &           61.4 &   55.6 &    58.4 &    56.8 &  \\textbf{64.9} &           63.3 \\\\\n",
      " A$\\rightarrow$W &  54.0 &           56.6 &   56.1 &    52.7 &    54.8 &  \\textbf{66.4} &           65.3 \\\\\n",
      " D$\\rightarrow$A &  42.4 &           45.5 &   42.2 &    44.1 &    51.1 &           54.3 &  \\textbf{57.9} \\\\\n",
      " D$\\rightarrow$W &  90.9 &           93.3 &   91.7 &    91.7 &    94.7 &             95 &  \\textbf{95.7} \\\\\n",
      " W$\\rightarrow$A &  40.8 &           45.5 &   41.0 &    42.7 &    49.7 &           52.7 &  \\textbf{56.2} \\\\\n",
      " W$\\rightarrow$D &  97.8 &  \\textbf{98.8} &   98.0 &    95.8 &    97.2 &           97.2 &           97.8 \\\\\n",
      "             AVG &  64.3 &           66.9 &   64.1 &    64.2 &    67.4 &           71.7 &  \\textbf{72.7} \\\\\n",
      "              nb &   2.0 &           1000 &    2.0 &  1000.0 &  1000.0 &           1000 &            286 \\\\\n",
      "\\bottomrule\n",
      "\\end{tabular}\n",
      "\n",
      "\\newpage Feature :$ office31fc7_SA $ Cheat  False \\newline\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\\begin{tabular}{lrlrrlll}\n",
      "\\toprule\n",
      "         dataset &    NA &             SA &  CORAL &     TCA &             OT &           OTDA &           MLOT \\\\\n",
      "\\midrule\n",
      " A$\\rightarrow$D &  59.6 &  \\textbf{61.0} &   55.6 &    58.0 &           56.8 &           55.4 &           54.8 \\\\\n",
      " A$\\rightarrow$W &  54.0 &  \\textbf{56.4} &   56.1 &    52.7 &           49.2 &           50.8 &           53.7 \\\\\n",
      " D$\\rightarrow$A &  42.4 &             45 &   42.2 &    44.0 &           48.8 &           48.1 &  \\textbf{50.4} \\\\\n",
      " D$\\rightarrow$W &  90.9 &           93.2 &   91.7 &    91.7 &           92.1 &  \\textbf{95.0} &           92.6 \\\\\n",
      " W$\\rightarrow$A &  40.8 &           45.1 &   41.0 &    42.5 &  \\textbf{48.3} &           46.9 &           47.6 \\\\\n",
      " W$\\rightarrow$D &  97.8 &  \\textbf{98.6} &   98.0 &    95.6 &           96.8 &           95.4 &           97.8 \\\\\n",
      "             AVG &  64.3 &  \\textbf{66.5} &   64.1 &    64.1 &           65.3 &           65.3 &           66.2 \\\\\n",
      "              nb &   2.0 &           1000 &    2.0 &  1000.0 &           1000 &           1000 &            286 \\\\\n",
      "\\bottomrule\n",
      "\\end{tabular}\n",
      "\n"
     ]
    }
   ],
   "source": [
    "recreate_table.creat_all_cheat_tables(\"NA,SA,CORAL,TCA,OT,OTDA,MLOT\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Images"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The table can be recreate only if the speficic_comparaison has already been launch."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "!python main.py -n --specific_comparaison OTDAvsMLOT --adaptationAlgoUsed OTDA,MLOT --pickle_name pickle_specific/"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import recreate_table"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "recreate_table.display_image(comparaison=True, color_map='Greens', save_image=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "recreate_table.display_image(comparaison=False, color_map='YlOrRd', save_image=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Non-exhaustive comparison between MLOT and ML + OTDA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Namespace(SA=False, adaptationAlgoUsed='MLOT', cross_val=False, cuda=False, d=70, detach=False, featuresToUse='surf', lr=0.001, margin=10, max_inner_iter_grad=1, max_iter=10, new_space=True, numberIteration=10, pickle_name='test', reg_cl=0.2, reg_e=0.2, reg_l=1, reg_pca=1, rule='median', save_pickle=False, specific_comparaison='None', svd_solver='full', test_exp=False, time_cross_val=1, type_classifier='1NN', verbose=False, which_dataset=-1)\n",
      "A->C .\n",
      "     41.3  0.0  MLOT    10.40s\n",
      "A->D .\n",
      "     39.5  0.0  MLOT     6.49s\n",
      "A->W .\n",
      "     48.1  0.0  MLOT     6.59s\n",
      "C->A .\n",
      "     55.5  0.0  MLOT    11.41s\n",
      "C->D .\n",
      "     52.9  0.0  MLOT     8.34s\n",
      "C->W .\n",
      "     54.9  0.0  MLOT     8.50s\n",
      "D->A .\n",
      "     40.0  0.0  MLOT     4.53s\n",
      "D->C .\n",
      "     33.7  0.0  MLOT     4.33s\n",
      "D->W .\n",
      "     85.4  0.0  MLOT     4.29s\n",
      "W->A .\n",
      "     40.1  0.0  MLOT     6.22s\n",
      "W->C .\n",
      "     36.1  0.0  MLOT    14.22s\n",
      "W->D .\n",
      "     75.2  0.0  MLOT     8.13s\n",
      "\n",
      "Mean results and total time\n",
      "     50.2  0.0  MLOT    93.43s\n"
     ]
    }
   ],
   "source": [
    "!python main.py -n --adaptationAlgoUsed MLOT --svd_solver full"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Namespace(SA=False, adaptationAlgoUsed='MLOT', cross_val=False, cuda=False, d=70, detach=False, featuresToUse='surf', lr=0.001, margin=10, max_inner_iter_grad=1000, max_iter=2, new_space=True, numberIteration=10, pickle_name='test', reg_cl=0.2, reg_e=0.2, reg_l=0.0, reg_pca=1, rule='median', save_pickle=False, specific_comparaison='None', svd_solver='full', test_exp=False, time_cross_val=1, type_classifier='1NN', verbose=False, which_dataset=-1)\n",
      "A->C .\n",
      "     40.5  0.0  MLOT     7.24s\n",
      "A->D .\n",
      "     38.9  0.0  MLOT     4.50s\n",
      "A->W .\n",
      "     44.7  0.0  MLOT    21.79s\n",
      "C->A .\n",
      "     54.0  0.0  MLOT    15.11s\n",
      "C->D .\n",
      "     47.8  0.0  MLOT     7.52s\n",
      "C->W .\n",
      "     50.5  0.0  MLOT    18.17s\n",
      "D->A .\n",
      "     39.2  0.0  MLOT     4.56s\n",
      "D->C .\n",
      "     33.3  0.0  MLOT     4.68s\n",
      "D->W .\n",
      "     84.1  0.0  MLOT     7.75s\n",
      "W->A .\n",
      "     39.0  0.0  MLOT     2.17s\n",
      "W->C .\n",
      "     35.4  0.0  MLOT     8.03s\n",
      "W->D .\n",
      "     79.6  0.0  MLOT     3.69s\n",
      "\n",
      "Mean results and total time\n",
      "     48.9  0.0  MLOT   105.20s\n"
     ]
    }
   ],
   "source": [
    "!python main.py -n --adaptationAlgoUsed MLOT --svd_solver full --max_iter 2 --max_inner_iter_grad 1000 --reg_l 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Namespace(SA=False, adaptationAlgoUsed='MLOT', cross_val=False, cuda=False, d=70, detach=False, featuresToUse='surf', lr=0.001, margin=10, max_inner_iter_grad=1, max_iter=10, new_space=True, numberIteration=10, pickle_name='test', reg_cl=0.2, reg_e=0.2, reg_l=1, reg_pca=1, rule='median', save_pickle=False, specific_comparaison='None', svd_solver='auto', test_exp=False, time_cross_val=1, type_classifier='1NN', verbose=False, which_dataset=-1)\n",
      "A->C ..........\n",
      "     41.9  0.5  MLOT   573.42s\n",
      "A->D ..........\n",
      "     40.0  1.7  MLOT   374.13s\n",
      "A->W ..........\n",
      "     49.2  1.3  MLOT   393.84s\n",
      "C->A ..........\n",
      "     56.1  0.5  MLOT   644.97s\n",
      "C->D ..........\n",
      "     52.6  1.0  MLOT   449.56s\n",
      "C->W ..........\n",
      "     55.0  1.2  MLOT   461.70s\n",
      "D->A ..........\n",
      "     40.3  0.7  MLOT   225.84s\n",
      "D->C ..........\n",
      "     34.9  0.6  MLOT   235.67s\n",
      "D->W ..........\n",
      "     83.2  1.4  MLOT   217.49s\n",
      "W->A ..........\n",
      "     40.6  0.6  MLOT   299.21s\n",
      "W->C ..........\n",
      "     35.5  0.5  MLOT   282.59s\n",
      "W->D ..........\n",
      "     77.6  1.3  MLOT   226.46s\n",
      "\n",
      "Mean results and total time\n",
      "     50.6  0.9  MLOT   4384.89s\n"
     ]
    }
   ],
   "source": [
    "!python main.py -n --adaptationAlgoUsed MLOT"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Namespace(SA=False, adaptationAlgoUsed='MLOT', cross_val=False, cuda=False, d=70, detach=False, featuresToUse='surf', lr=0.001, margin=10, max_inner_iter_grad=1000, max_iter=2, new_space=True, numberIteration=10, pickle_name='test', reg_cl=0.2, reg_e=0.2, reg_l=0.0, reg_pca=1, rule='median', save_pickle=False, specific_comparaison='None', svd_solver='auto', test_exp=False, time_cross_val=1, type_classifier='1NN', verbose=False, which_dataset=-1)\n",
      "A->C ..........\n",
      "     40.5  0.3  MLOT   399.35s\n",
      "A->D ..........\n",
      "     40.4  1.5  MLOT   608.33s\n",
      "A->W ..........\n",
      "     45.5  1.4  MLOT   474.73s\n",
      "C->A ..........\n",
      "     54.5  0.6  MLOT   833.26s\n",
      "C->D ..........\n",
      "     49.1  1.9  MLOT   829.90s\n",
      "C->W ..........\n",
      "     51.3  1.1  MLOT   930.67s\n",
      "D->A ..........\n",
      "     39.1  0.3  MLOT   281.25s\n",
      "D->C ..........\n",
      "     33.8  0.5  MLOT   278.25s\n",
      "D->W ..........\n",
      "     84.6  0.7  MLOT   212.75s\n",
      "W->A ..........\n",
      "     39.2  0.6  MLOT   280.92s\n",
      "W->C ..........\n",
      "     35.3  0.4  MLOT   319.45s\n",
      "W->D ..........\n",
      "     77.1  1.3  MLOT   185.39s\n",
      "\n",
      "Mean results and total time\n",
      "     49.2  0.9  MLOT   5634.26s\n"
     ]
    }
   ],
   "source": [
    "!python main.py -n --adaptationAlgoUsed MLOT --max_iter 2 --max_inner_iter_grad 1000 --reg_l 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Namespace(SA=False, adaptationAlgoUsed='MLOT_id', cross_val=False, cuda=False, d=70, detach=False, featuresToUse='surf', lr=0.001, margin=10, max_inner_iter_grad=1000, max_iter=2, new_space=True, numberIteration=10, pickle_name='test', reg_cl=0.2, reg_e=0.2, reg_l=0.0, reg_pca=1, rule='median', save_pickle=False, specific_comparaison='None', svd_solver='auto', test_exp=False, time_cross_val=1, type_classifier='1NN', verbose=False, which_dataset=-1)\n",
      "A->C .\n",
      "     40.3  0.0  MLOT_id  10.80s\n",
      "A->D .\n",
      "     39.5  0.0  MLOT_id   6.55s\n",
      "A->W .\n",
      "     42.4  0.0  MLOT_id  16.37s\n",
      "C->A .\n",
      "     51.3  0.0  MLOT_id  30.20s\n",
      "C->D .\n",
      "     48.4  0.0  MLOT_id   6.69s\n",
      "C->W .\n",
      "     49.2  0.0  MLOT_id  21.43s\n",
      "D->A .\n",
      "     36.7  0.0  MLOT_id   1.86s\n",
      "D->C .\n",
      "     35.4  0.0  MLOT_id   1.98s\n",
      "D->W .\n",
      "     82.4  0.0  MLOT_id   1.60s\n",
      "W->A .\n",
      "     39.6  0.0  MLOT_id   2.72s\n",
      "W->C .\n",
      "     35.7  0.0  MLOT_id   2.58s\n",
      "W->D .\n",
      "     75.2  0.0  MLOT_id   3.47s\n",
      "\n",
      "Mean results and total time\n",
      "     48.0  0.0  MLOT_id 106.23s\n"
     ]
    }
   ],
   "source": [
    "!python main.py -n --adaptationAlgoUsed MLOT_id --max_iter 2 --max_inner_iter_grad 1000 --reg_l 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Namespace(SA=False, adaptationAlgoUsed='MLOT_id', cross_val=False, cuda=False, d=70, detach=False, featuresToUse='surf', lr=0.001, margin=10, max_inner_iter_grad=1, max_iter=10, new_space=True, numberIteration=10, pickle_name='test', reg_cl=0.2, reg_e=0.2, reg_l=1, reg_pca=1, rule='median', save_pickle=False, specific_comparaison='None', svd_solver='auto', test_exp=False, time_cross_val=1, type_classifier='1NN', verbose=False, which_dataset=-1)\n",
      "A->C .\n",
      "     40.6  0.0  MLOT_id  11.29s\n",
      "A->D .\n",
      "     42.0  0.0  MLOT_id   7.53s\n",
      "A->W .\n",
      "     43.4  0.0  MLOT_id   7.85s\n",
      "C->A .\n",
      "     51.7  0.0  MLOT_id  11.83s\n",
      "C->D .\n",
      "     46.5  0.0  MLOT_id   8.45s\n",
      "C->W .\n",
      "     49.8  0.0  MLOT_id   8.79s\n",
      "D->A .\n",
      "     38.1  0.0  MLOT_id   4.33s\n",
      "D->C .\n",
      "     34.8  0.0  MLOT_id   5.20s\n",
      "D->W .\n",
      "     81.7  0.0  MLOT_id   3.90s\n",
      "W->A .\n",
      "     38.1  0.0  MLOT_id   4.84s\n",
      "W->C .\n",
      "     35.1  0.0  MLOT_id   5.50s\n",
      "W->D .\n",
      "     75.2  0.0  MLOT_id   3.96s\n",
      "\n",
      "Mean results and total time\n",
      "     48.1  0.0  MLOT_id  83.47s\n"
     ]
    }
   ],
   "source": [
    "!python main.py -n --adaptationAlgoUsed MLOT_id"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Pytorch version of MLOT"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "OTSAMLl and OTSAMLnl are two implementation of MLOT but instead of using LMNN update, we use pytorch automatic differentiation. This explain the little difference between the OTSAMLl (which has only a linear matrix as NN) and OTML. OTSAMLnl is the non-linear version, the exact architecture of this NN is available in the code. This code is still in progress and it is clearly not optimize as there is very often swap between numpy and torch tensor."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "import recreate_table"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\\newpage Feature :$ surf_SA $ Cheat  True \\newline\n",
      "\\begin{tabular}{lll}\n",
      "\\toprule\n",
      "         dataset &        OTSAMLl &       OTSAMLnl \\\\\n",
      "\\midrule\n",
      " A$\\rightarrow$C &  \\textbf{43.1} &           42.9 \\\\\n",
      " A$\\rightarrow$D &           45.9 &  \\textbf{49.0} \\\\\n",
      " A$\\rightarrow$W &  \\textbf{52.9} &           52.5 \\\\\n",
      " C$\\rightarrow$A &           53.4 &  \\textbf{55.7} \\\\\n",
      " C$\\rightarrow$D &           51.6 &  \\textbf{55.4} \\\\\n",
      " C$\\rightarrow$W &           56.6 &  \\textbf{56.9} \\\\\n",
      " D$\\rightarrow$A &  \\textbf{43.2} &           40.9 \\\\\n",
      " D$\\rightarrow$C &           36.7 &  \\textbf{37.1} \\\\\n",
      " D$\\rightarrow$W &  \\textbf{90.8} &           90.8 \\\\\n",
      " W$\\rightarrow$A &  \\textbf{44.5} &           43.8 \\\\\n",
      " W$\\rightarrow$C &           36.6 &  \\textbf{37.2} \\\\\n",
      " W$\\rightarrow$D &  \\textbf{92.4} &           92.4 \\\\\n",
      "             AVG &             54 &  \\textbf{54.6} \\\\\n",
      "              nb &          366.8 &          498.7 \\\\\n",
      "\\bottomrule\n",
      "\\end{tabular}\n",
      "\n",
      "\\newpage Feature :$ surf_SA $ Cheat  False \\newline\n",
      "\\begin{tabular}{lll}\n",
      "\\toprule\n",
      "         dataset &        OTSAMLl &       OTSAMLnl \\\\\n",
      "\\midrule\n",
      " A$\\rightarrow$C &  \\textbf{40.5} &           38.9 \\\\\n",
      " A$\\rightarrow$D &           39.5 &  \\textbf{40.1} \\\\\n",
      " A$\\rightarrow$W &  \\textbf{42.7} &             40 \\\\\n",
      " C$\\rightarrow$A &  \\textbf{53.1} &           52.9 \\\\\n",
      " C$\\rightarrow$D &  \\textbf{51.6} &             49 \\\\\n",
      " C$\\rightarrow$W &  \\textbf{43.1} &           43.1 \\\\\n",
      " D$\\rightarrow$A &           36.6 &  \\textbf{37.3} \\\\\n",
      " D$\\rightarrow$C &           34.3 &  \\textbf{34.4} \\\\\n",
      " D$\\rightarrow$W &  \\textbf{89.2} &           88.1 \\\\\n",
      " W$\\rightarrow$A &           34.8 &  \\textbf{36.1} \\\\\n",
      " W$\\rightarrow$C &           31.8 &  \\textbf{33.6} \\\\\n",
      " W$\\rightarrow$D &           89.8 &  \\textbf{91.1} \\\\\n",
      "             AVG &  \\textbf{48.9} &           48.7 \\\\\n",
      "              nb &          366.8 &          498.7 \\\\\n",
      "\\bottomrule\n",
      "\\end{tabular}\n",
      "\n"
     ]
    }
   ],
   "source": [
    "recreate_table.creat_all_cheat_tables(names=\"OTSAMLl,OTSAMLnl\", features=\"surf_SA\",)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}