graphs_kernels_pipeline_demo.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Graph kernels with PASQAL processor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "from utils import jensen_shannon, return_fourier_from_dataset\n",
    "from data_preprocessing import load_dataset\n",
    "import pickle\n",
    "import networkx as nx\n",
    "import grakel\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn import svm\n",
    "from sklearn.metrics import accuracy_score, f1_score, make_scorer\n",
    "from sklearn.model_selection import cross_val_score"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Import the IMDB-MULTI dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We import the IMDB-MULTI dataset, limited to the graphs below 10 nodes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of graphs:  814\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "graphs, targets = load_dataset('IMDB-MULTI', min_node=0, max_node=10)\n",
    "print(\"Number of graphs: \", len(graphs))\n",
    "\n",
    "plt.figure(figsize=[12,6])\n",
    "plt.subplot(121)\n",
    "nx.draw(graphs[0], with_labels=True)\n",
    "plt.title(\"Graph 0\", fontsize='xx-large')\n",
    "plt.subplot(122)\n",
    "nx.draw(graphs[1], with_labels=True)\n",
    "plt.title(\"Graph 1\", fontsize='xx-large')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For the rest of the demo, we will limit ourselves to a sample of 20 graphs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.random.seed(40)\n",
    "sample = np.random.choice(len(graphs), size=9, replace=False).astype(int)\n",
    "\n",
    "graphs_sample, targets_sample = graphs[sample], targets[sample]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Generate the hamiltonian evolutions"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Generate the fourier signals"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/constantindalyac/Desktop/Pasqal/qgraph/utils.py:178: ComplexWarning: Casting complex values to real discards the imaginary part\n",
      "  hamiltonian='xy')\n",
      "/Users/constantindalyac/Desktop/Pasqal/qgraph/utils.py:182: ComplexWarning: Casting complex values to real discards the imaginary part\n",
      "  hamiltonian='ising')\n"
     ]
    }
   ],
   "source": [
    "#Relabel the nodes\n",
    "graphs_sample = [nx.convert_node_labels_to_integers(G) for G in graphs_sample]\n",
    "graphs_sample = np.array(graphs_sample, dtype=object)\n",
    "\n",
    "#Generate signals\n",
    "fourier_xy, fourier_ising = return_fourier_from_dataset(graphs_sample)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Discretize into histograms"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "histograms_xy = []\n",
    "histograms_ising = []\n",
    "\n",
    "for i in range(len(graphs_sample)):\n",
    "    histograms_xy.append(np.histogram(fourier_xy[1,i], bins=500, weights=np.abs(fourier_xy[0,i])**2, density=True))\n",
    "    histograms_ising.append(np.histogram(fourier_ising[1,i], bins=500, weights=np.abs(fourier_ising[0,i])**2, density=True))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Plot some histograms of XY evolution"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=[12,6])\n",
    "plt.subplot(121)\n",
    "plt.bar(histograms_xy[0][1][1::], histograms_xy[0][0])\n",
    "plt.title(\"Graph 0\", fontsize='xx-large')\n",
    "plt.subplot(122)\n",
    "plt.bar(histograms_xy[1][1][1::], histograms_xy[1][0])\n",
    "plt.title(\"Graph 1\", fontsize='xx-large')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Compute the distance matrices"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Initialize empty matrices\n",
    "js_xy = np.zeros((len(graphs_sample), len(graphs_sample)))\n",
    "js_ising = np.zeros((len(graphs_sample), len(graphs_sample)))\n",
    "\n",
    "for i in range(len(graphs_sample)):\n",
    "    for j in range(i, len(graphs_sample)):\n",
    "        js_xy[i,j] = jensen_shannon(histograms_xy[i], histograms_xy[j])\n",
    "        js_ising[i,j] = jensen_shannon(histograms_ising[i], histograms_ising[j])\n",
    "        js_xy[j,i] = js_xy[i,j]\n",
    "        js_ising[j,i] = js_ising[i,j]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Compute the classical kernels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Turn the graphs into grakel.Graph objects\n",
    "graphs_grakel = [grakel.Graph(nx.adjacency_matrix(graph)) for graph in graphs_sample]\n",
    "\n",
    "rw_kernel = grakel.RandomWalk(lamda=0.001)\n",
    "gs_kernel = grakel.GraphletSampling(k=6, sampling={'n_samples':100})\n",
    "lt_kernel = grakel.LovaszTheta()\n",
    "\n",
    "K_rw = rw_kernel.fit_transform(graphs_grakel)\n",
    "K_gs = gs_kernel.fit_transform(graphs_grakel)\n",
    "K_lt = lt_kernel.fit_transform(graphs_grakel)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Fit the models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Mean of F1 score for quantum kernel:  0.512962962962963\n",
      "Mean of F1 score for RW kernel:  0.36296296296296293\n",
      "Mean of F1 score for GS kernel:  0.3333333333333333\n",
      "Mean of F1 score for LT kernel:  0.3833333333333333\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Applications/anaconda3/lib/python3.7/site-packages/sklearn/model_selection/_split.py:657: Warning: The least populated class in y has only 2 members, which is too few. The minimum number of members in any class cannot be less than n_splits=3.\n",
      "  % (min_groups, self.n_splits)), Warning)\n",
      "/Applications/anaconda3/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Applications/anaconda3/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Applications/anaconda3/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Applications/anaconda3/lib/python3.7/site-packages/sklearn/model_selection/_split.py:657: Warning: The least populated class in y has only 2 members, which is too few. The minimum number of members in any class cannot be less than n_splits=3.\n",
      "  % (min_groups, self.n_splits)), Warning)\n",
      "/Applications/anaconda3/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Applications/anaconda3/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Applications/anaconda3/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Applications/anaconda3/lib/python3.7/site-packages/sklearn/model_selection/_split.py:657: Warning: The least populated class in y has only 2 members, which is too few. The minimum number of members in any class cannot be less than n_splits=3.\n",
      "  % (min_groups, self.n_splits)), Warning)\n",
      "/Applications/anaconda3/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Applications/anaconda3/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Applications/anaconda3/lib/python3.7/site-packages/sklearn/model_selection/_split.py:657: Warning: The least populated class in y has only 2 members, which is too few. The minimum number of members in any class cannot be less than n_splits=3.\n",
      "  % (min_groups, self.n_splits)), Warning)\n",
      "/Applications/anaconda3/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Applications/anaconda3/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n",
      "/Applications/anaconda3/lib/python3.7/site-packages/sklearn/metrics/classification.py:1437: UndefinedMetricWarning: F-score is ill-defined and being set to 0.0 in labels with no predicted samples.\n",
      "  'precision', 'predicted', average, warn_for)\n"
     ]
    }
   ],
   "source": [
    "p = 0.2\n",
    "mu = 10\n",
    "C = 10\n",
    "\n",
    "#Compute the quantum kernel from the distances matrices\n",
    "K_q = np.exp(-mu * (p * js_xy + (1-p) * js_ising))\n",
    "\n",
    "#Initialize the classifier\n",
    "clf = svm.SVC(kernel='precomputed', C=C, random_state=76)\n",
    "\n",
    "#Perform a 3-fold cross-validation for every kernel\n",
    "scores_q = cross_val_score(clf, K_q, targets_sample, cv=3, scoring=make_scorer(f1_score, average='weighted'))\n",
    "scores_rw = cross_val_score(clf, K_rw, targets_sample, cv=3, scoring=make_scorer(f1_score, average='weighted'))\n",
    "scores_gs = cross_val_score(clf, K_gs, targets_sample, cv=3, scoring=make_scorer(f1_score, average='weighted'))\n",
    "scores_lt = cross_val_score(clf, K_lt, targets_sample, cv=3, scoring=make_scorer(f1_score, average='weighted'))\n",
    "\n",
    "print(\"Mean of F1 score for quantum kernel: \", np.mean(scores_q))\n",
    "print(\"Mean of F1 score for RW kernel: \", np.mean(scores_rw))\n",
    "print(\"Mean of F1 score for GS kernel: \", np.mean(scores_gs))\n",
    "print(\"Mean of F1 score for LT kernel: \", np.mean(scores_lt))\n",
    "\n"
   ]
  },
  {
   "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}