MVA-2021 / graphs_in_ml / hw2_ssl / SSL.ipynb
SSL.ipynb
Raw
{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Semi-Supervised Learning (SSL)\n",
    "\n",
    "\n",
    "SSL studies how to learn from both labeled and unlabeled data, which can be useful when data is abundant but the resources to label them are limited.\n",
    "\n",
    "In this exercise, you will:\n",
    "\n",
    "* Given a simulated dataset with both labeled and unlabeled data, build a similarity graph and use the Harmonic Function Solution (HSF) to predict the labels of the unlabeled data;\n",
    "* Use HSF for face recognition, given a fixed dataset;\n",
    "* Implement an online version of HSF to label images as they appear in real time."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. Harmonic Function Solution\n",
    "\n",
    "Let $G = (V, E)$ be a weighted undirected graph where $V = \\{x_1, \\ldots, x_n \\}$ is the vertex set and $E$ is the edge set. Each edge $e_{ij} \\in E$ has a weight $w_{ij}$ and, if there is no edge between $x_i$ and $x_j$, then $w_{ij}=0$.\n",
    "\n",
    "Let $|V| = n$ be the total number of nodes. Only a subset of the nodes $S \\subset V$ with cardinality $|S| = l$ is labeled, and the remaining $u = n - l$ nodes are placed in the subset $T = V \\setminus S$. \n",
    "\n",
    "Our goal is to predict the labels of the vertices in $T$ using the structure of the graph. Since we believe that nodes close in the graph should have similar labels, we would like to have each node surrounded by a majority of nodes with the same label. In order to do so, we impose that the labeling vector $f \\in \\mathbb{R}^n$ must be an **harmonic function** on the graph, that is:\n",
    "\n",
    "$$\n",
    "f_i = \\frac{\\sum_{j} w_{ij} f_j}{\\sum_{j} w_{ij}},  \\forall i \\in T\n",
    "$$\n",
    "\n",
    "One interpretation for this constraint is that $w_{ij}$ represents the tendency of moving from node $x_i$ to node $x_j$, the stationary distribution of the transition matrix $P(j|i) = \\tfrac{w_{ij}}{\\sum_{k} w_{ik}}$  is a valid solution to our problem. \n",
    "\n",
    "### Hard HFS\n",
    "\n",
    "It can be shown that $f$ is harmonic if and only if $(Lf)_T = 0$, where $(Lf)_T$ is the vector containing the values of $Lf$ for the nodes in the set $T$, and $L$ is the graph Laplacian. \n",
    "\n",
    "Hence, the harmonic function solution to the SSL problem is the solution to the following optimization problem:\n",
    "\n",
    "$$\n",
    "\\min_{f \\in \\mathbb{R}^n}  f^T L f  \n",
    "\\quad \\text{s.t} \\quad\n",
    "y_i = f(x_i) \\quad \\forall x_i \\in S\n",
    "$$\n",
    "where $y_i$ are the labels available for the vertices $x_i \\in S$. This gives us:\n",
    "\n",
    "$$\n",
    "f_T = L_{TT}^{-1}(W_{TS}f_S)\n",
    "$$\n",
    "\n",
    "### Soft HFS\n",
    "\n",
    "If the labels are noisy, we might need to replace the \"hard\" constraint of the optimization problem above by a \"soft\" constraint. Let $C$ be a diagonal matrix such that $C_{ii} = c_l$ for labeled examples and $C_{ii} = c_u$ otherwise. Also, define $y_i = 0$ for unlabeled examples, that is, for $x_i \\in T$. \n",
    "\n",
    "The soft HFS objective function is\n",
    "\n",
    "$$\n",
    "\\min_{f\\in\\mathbb{R}^n} (f-y)^T C (f-y) + f^T L f\n",
    "$$\n",
    "whose solution is \n",
    "\n",
    "$$\n",
    "f^* = (C^{-1}L+I)^{-1}y\n",
    "$$\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Implementation\n",
    "\n",
    "Implement hard and soft HFS in the function `compute_hfs`. Complete the function `two_moons_hfs` to test your implementation using the datasets `data_2moons_hfs.mat` and `data_2moons_hfs_large.mat`.\n",
    "\n",
    "\n",
    "* Tips: \n",
    "    * Don't forget to choose well the parameters to build the graph and its Laplacian.\n",
    "    * You can use the functions `build_laplacian_regularized` and `build_similarity_graph`. The function `mask_labels` is used to chose how many labels are revealed.\n",
    "    * Be careful: the labels are revealed randomly, and each random realization can have different results! Check how the `seed` parameter works.\n",
    "    * Introduce noisy labels to compare hard and soft HFS."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import scipy.spatial.distance as sd\n",
    "from scipy.io import loadmat\n",
    "import os\n",
    "from helper import build_similarity_graph, label_noise\n",
    "from helper import build_laplacian, build_laplacian_regularized\n",
    "from helper import plot_classification\n",
    "from helper import mask_labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "Define parameters for HFS\n",
    "\"\"\"\n",
    "params = {}\n",
    "\n",
    "# regularization parameter (gamma)\n",
    "params['laplacian_regularization'] = 0.0\n",
    "\n",
    "# the sigma value for the exponential (similarity) function, already squared\n",
    "params['var'] = 1.0\n",
    "\n",
    "# Threshold eps for epsilon graphs\n",
    "params['eps'] = None\n",
    "\n",
    "# Number of neighbours k for k-nn. If zero, use epsilon-graph\n",
    "params['k'] = None\n",
    "\n",
    "# String selecting which version of the laplacian matrix to construct.\n",
    "# 'unn':  unnormalized, 'sym': symmetric normalization, 'rw':  random-walk normalization \n",
    "params['laplacian_normalization'] = 'unn'\n",
    "\n",
    "# Coefficients for C matrix for soft HFS\n",
    "params['c_l'] = None\n",
    "params['c_u'] = None"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "def compute_hfs(L, Y, soft=False, **params):\n",
    "    \"\"\"\n",
    "    Function to perform hard (constrained) HFS.\n",
    "\n",
    "    Parameters\n",
    "    ----------\n",
    "    L : array\n",
    "        Graph Laplacian, (n x n) matrix (regularized or not)\n",
    "    Y : array\n",
    "        (n, ) array with nodes labels [0, 1, ... , num_classes] (0 is unlabeled)\n",
    "    soft : bool\n",
    "        If True, compute soft HFS. Otherwise, compute hard HFS.\n",
    "\n",
    "    Returns\n",
    "    --------\n",
    "        Labels, class assignments for each of the n nodes\n",
    "    \"\"\"\n",
    "\n",
    "    num_samples = L.shape[0]\n",
    "    Cl = np.unique(Y)\n",
    "    num_classes = len(Cl)-1\n",
    "\n",
    "    \"\"\"\n",
    "    Build the vectors:\n",
    "    y = (n x num_classes) target vector \n",
    "    l_idx = shape (l,) vector with indices of labeled nodes\n",
    "    u_idx = shape (u,) vector with indices of unlabeled nodes\n",
    "    \"\"\"\n",
    "    y = np.zeros((num_samples, num_classes+1))\n",
    "    y[np.arange(num_samples), Y.astype(int)] = 1 # onehot encoding\n",
    "    y = y[:,1:] # removing unlabeled column\n",
    "    l_idx = np.where(Y != 0)[0]\n",
    "    u_idx = np.where(Y == 0)[0]\n",
    "    \n",
    "    if not soft:    \n",
    "        \"\"\"\n",
    "        Compute hard HFS.  \n",
    "\n",
    "        f_l = solution for labeled data. \n",
    "        f_u = solution for unlabeled data\n",
    "        f   = solution for all data\n",
    "        \"\"\"\n",
    "        # onehot solution for labeled data is immediate\n",
    "        f_l = y[l_idx]  \n",
    "        # compute solution for unlabeled data\n",
    "        Luu = L[u_idx][:, u_idx]\n",
    "        Wul = -L[u_idx][:, l_idx]\n",
    "        # f_u = np.linalg.pinv(Luu) @ Wul @ f_l\n",
    "        f_u = np.linalg.solve(Luu, Wul @ f_l)\n",
    "        # create solution for all data\n",
    "        f = np.zeros((num_samples, num_classes))\n",
    "        f[l_idx] = f_l\n",
    "        f[u_idx] = f_u\n",
    "        \n",
    "\n",
    "    else:\n",
    "        \"\"\"\n",
    "        Compute soft HFS.\n",
    "        f = harmonic function solution \n",
    "        C = (n x n) diagonal matrix with c_l for labeled samples and c_u otherwise    \n",
    "        \"\"\"\n",
    "        # compute C\n",
    "        C = np.zeros(num_samples)\n",
    "        C[l_idx] = params['c_l']\n",
    "        C[u_idx] = params['c_u']\n",
    "        C = np.diag(C)\n",
    "        # compute f\n",
    "        # f = np.linalg.pinv(np.linalg.pinv(C) @ L + np.eye(num_samples)) @ y\n",
    "        f = np.linalg.solve(np.linalg.pinv(C) @ L + np.eye(num_samples), y)\n",
    "\n",
    "    \n",
    "    \"\"\"\n",
    "    return the labels assignment from the hfs solution, and the solution f\n",
    "    labels: (n x 1) class assignments [1,2,...,num_classes]    \n",
    "    f : harmonic function solution\n",
    "    \"\"\"\n",
    "    labels = np.argmax(f, axis=1) + 1\n",
    "    \n",
    "    return labels, f\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 169,
   "metadata": {},
   "outputs": [],
   "source": [
    "def two_moons_hfs(l=4, l_noisy=1, soft=False, dataset='data_2moons_hfs.mat', plot=True, seed=None, **params):\n",
    "    \"\"\" \n",
    "    HFS for two_moons data.\n",
    "    \n",
    "    Parameters\n",
    "    ----------\n",
    "    l : int\n",
    "        Number of labeled (unmasked) nodes provided to the HFS algorithm.\n",
    "    l_noisy : int\n",
    "        Number of *noisy* labels to introduce.\n",
    "    soft : bool\n",
    "        If true, use soft HFS, otherwise use hard HFS\n",
    "    dataset : {'data_2moons_hfs.mat' or 'data_2moons_hfs_large.mat'}\n",
    "        Which dataset to use.\n",
    "    plot : bool\n",
    "        If True, show plots\n",
    "    seed : int\n",
    "        If not None, set global numpy seed before choosing labels to reveal.\n",
    "    \"\"\"\n",
    "    if seed is not None:\n",
    "        np.random.seed(seed)\n",
    "\n",
    "    # Load the data. At home, try to use the larger dataset.    \n",
    "    in_data = loadmat(os.path.join('data', dataset))\n",
    "    X = in_data['X']\n",
    "    Y = np.array(in_data['Y'].squeeze(), dtype=np.uint32)\n",
    "\n",
    "    # infer number of labels from samples\n",
    "    num_samples = np.size(Y, 0)\n",
    "    unique_classes = np.unique(Y)\n",
    "    num_classes = len(unique_classes)\n",
    "    \n",
    "    # mask labels\n",
    "    Y_masked = mask_labels(Y, l)\n",
    "    # Y_masked = mask_labels(Y, int(l / num_classes), per_class=True)\n",
    "    assert len(np.unique(Y_masked)) > 2, \"only one class in training data!\"\n",
    "    # introduce noise\n",
    "    noise_indices = np.where(Y_masked == 0)[0]\n",
    "    np.random.shuffle(noise_indices)\n",
    "    noise_indices = noise_indices[:l_noisy]\n",
    "    Y_masked[noise_indices] = np.random.choice(unique_classes, l_noisy)\n",
    "\n",
    "    \"\"\"\n",
    "    compute hfs solution using either soft_hfs or hard_hfs\n",
    "    \"\"\"\n",
    "    # Build graph Laplacian using the parameters:\n",
    "    # params['laplacian_regularization'], params['var'], params['eps'], \n",
    "    # params['k'] and params['laplacian_normalization'].\n",
    "    \n",
    "    L = build_laplacian_regularized(X, params['laplacian_regularization'], params['var'], \n",
    "                                    params['eps'], params['k'], params['laplacian_normalization'])\n",
    "\n",
    "    labels, f = compute_hfs(L, Y_masked, soft, **params)\n",
    "\n",
    "    # Visualize results\n",
    "    if plot:\n",
    "        plot_classification(X, Y, Y_masked, noise_indices, labels, params['var'], params['eps'], params['k'])\n",
    "    accuracy = np.mean(labels == np.squeeze(Y))\n",
    "    print(f\"Soft={soft}, Accuracy={accuracy}\")\n",
    "    return X, Y, labels, accuracy"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Question 1.1 - Report the accuracy you obtained for `data_2moons_hfs.mat` dataset using hard HFS, when l=10 and l_noisy=0. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x360 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Soft=False, Accuracy=1.0\n"
     ]
    }
   ],
   "source": [
    "seed = 42\n",
    "# the sigma value for the exponential (similarity) function, already squared\n",
    "params['var'] = 1.0\n",
    "# Threshold eps for epsilon graphs\n",
    "params['eps'] = 0.5\n",
    "# Number of neighbours k for k-nn. If zero, use epsilon-graph\n",
    "params['k'] = 0\n",
    "\n",
    "X, Y, hard_labels, hard_accuracy = two_moons_hfs(l=10, l_noisy=0, soft=False, dataset='data_2moons_hfs.mat',\n",
    "                                                 plot=True, seed=seed, **params)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Question 1.2  - Using `data_2moons_hfs_large.mat`, run `two_moons_hfs` several times with l=4. What can go wrong?\n",
    "\n",
    "* Tips:\n",
    "    * When running `two_moons_hfs` several times, don't forget to set `seed=None`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 171,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Soft=False, Accuracy=1.0\n",
      "Soft=False, Accuracy=1.0\n",
      "Soft=False, Accuracy=1.0\n",
      "Soft=False, Accuracy=1.0\n",
      "Soft=False, Accuracy=1.0\n",
      "Soft=False, Accuracy=1.0\n",
      "Soft=False, Accuracy=1.0\n",
      "Soft=False, Accuracy=1.0\n",
      "Soft=False, Accuracy=1.0\n",
      "Soft=False, Accuracy=1.0\n"
     ]
    },
    {
     "ename": "AssertionError",
     "evalue": "only one class in training data!",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mAssertionError\u001b[0m                            Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-171-2bbe9d63986a>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[0;32m     10\u001b[0m     X, Y, hard_labels, hard_accuracy = two_moons_hfs(l=4, l_noisy=0, soft=False, \n\u001b[0;32m     11\u001b[0m                                                      \u001b[0mdataset\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;34m'data_2moons_hfs_large.mat'\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 12\u001b[1;33m                                                      plot=False, seed=None, **params)\n\u001b[0m",
      "\u001b[1;32m<ipython-input-169-8aeb19c3dd66>\u001b[0m in \u001b[0;36mtwo_moons_hfs\u001b[1;34m(l, l_noisy, soft, dataset, plot, seed, **params)\u001b[0m\n\u001b[0;32m     34\u001b[0m     \u001b[0mY_masked\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mmask_labels\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mY\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0ml\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     35\u001b[0m     \u001b[1;31m# Y_masked = mask_labels(Y, int(l / num_classes), per_class=True)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 36\u001b[1;33m     \u001b[1;32massert\u001b[0m \u001b[0mlen\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mnp\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0munique\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mY_masked\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m \u001b[1;33m>\u001b[0m \u001b[1;36m2\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;34m\"only one class in training data!\"\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m     37\u001b[0m     \u001b[1;31m# introduce noise\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     38\u001b[0m     \u001b[0mnoise_indices\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mwhere\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mY_masked\u001b[0m \u001b[1;33m==\u001b[0m \u001b[1;36m0\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mAssertionError\u001b[0m: only one class in training data!"
     ]
    }
   ],
   "source": [
    "# the sigma value for the exponential (similarity) function, already squared\n",
    "params['var'] = 1.0\n",
    "# Threshold eps for epsilon graphs\n",
    "params['eps'] = 0.5\n",
    "# Number of neighbours k for k-nn. If zero, use epsilon-graph\n",
    "params['k'] = 0\n",
    "\n",
    "# np.random.seed(42)\n",
    "for ii in range(20):\n",
    "    X, Y, hard_labels, hard_accuracy = two_moons_hfs(l=4, l_noisy=0, soft=False, \n",
    "                                                     dataset='data_2moons_hfs_large.mat',\n",
    "                                                     plot=False, seed=None, **params)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "When running multiple times for $l=4$, i.e. sampling $4$ labeled points, most of the time the HFS gives an accuracy of $1$ for the clustering. However, it happens that the $4$ labeled points that are sampled come from the same label. In that case, the HFS can not solve the clustering problem.\n",
    "\n",
    "For this toy dataset, the issue can be resolved by using ```per_class = True``` in the ```mask_label``` function."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Question 1.3 - Using `data_2moons_hfs.mat`, l=10 and l_noisy=5, compare hard HFS to soft HFS. Report the accuracy and comment the results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x360 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Soft=False, Accuracy=0.885\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x360 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Soft=True, Accuracy=0.93\n"
     ]
    }
   ],
   "source": [
    "# the sigma value for the exponential (similarity) function, already squared\n",
    "params['var'] = 1.0\n",
    "# Threshold eps for epsilon graphs\n",
    "params['eps'] = 0.5\n",
    "# Number of neighbours k for k-nn. If zero, use epsilon-graph\n",
    "params['k'] = 0\n",
    "# Coefficients for C matrix for soft HFS\n",
    "params['c_l'] = 1\n",
    "params['c_u'] = 0.1\n",
    "\n",
    "# Comparing\n",
    "seed = 5  # To run several times with random outcomes, set seed=None. Otherwise, set a seed for reproducibility.\n",
    "plot = True \n",
    "dataset = 'data_2moons_hfs.mat' # Try also 'data_2moons_hfs_large.mat'\n",
    "\n",
    "X, Y, hard_labels, hard_accuracy = two_moons_hfs(l=10, l_noisy=5, soft=False, dataset=dataset,\n",
    "                                                 plot=plot, seed=seed, **params)\n",
    "X, Y, soft_labels, soft_accuracy = two_moons_hfs(l=10, l_noisy=5, soft=True, dataset=dataset,\n",
    "                                                 plot=plot, seed=seed, **params)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Based on the above example, we see that in the case of noisy labels the hard HFS tends to give more importance to wrong labels than soft HFS, which can be an issue. In the above example, it is typically the case for the labeled point at around $(x=-1.5, y=1)$ that supposedly had been given an alternative label. It results for the hard HFS in a small area where the points are not well-assigned, while the soft area can deal with the issue thanks to the well-assigned labeled points that are not far away from the wrongly assigned one. Therefore the final accuracy of soft HFS is better than the one of hard HFS ($0.93$ against $0.885$).\n",
    "\n",
    "We can repeat the experiment for different seeds, and see if it is still the case in average:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Comparing\n",
    "plot = False\n",
    "\n",
    "N = 100\n",
    "\n",
    "hard_accuracies = np.zeros(N)\n",
    "soft_accuracies = np.zeros(N)\n",
    "for seed in range(N): \n",
    "    dataset = 'data_2moons_hfs.mat' # Try also 'data_2moons_hfs_large.mat'\n",
    "\n",
    "    X, Y, hard_labels, hard_accuracy = two_moons_hfs(l=10, l_noisy=5, soft=False, dataset=dataset,\n",
    "                                                     plot=plot, seed=seed, **params)\n",
    "    X, Y, soft_labels, soft_accuracy = two_moons_hfs(l=10, l_noisy=5, soft=True, dataset=dataset,\n",
    "                                                     plot=plot, seed=seed, **params)\n",
    "    hard_accuracies[seed] = hard_accuracy\n",
    "    soft_accuracies[seed] = soft_accuracy  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "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": [
    "plt.hist(hard_accuracies, bins=100, label=\"Hard\", alpha=0.5)\n",
    "plt.hist(soft_accuracies, bins=100, label=\"Soft\", alpha=0.5)\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Above is the histogram of the accuracies for the first $100$ seeds. It shows that soft HFS behaves better than hard HFS in this case."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. Face recognition with HFS\n",
    "\n",
    "Now, we apply HFS to the task of face recognition, that is, our goal is to classify faces as belonging to different people. Since faces all share common features, it can be a good idea to leverage a large quantity of unlabeled data to improve classification accuracy. In this part of the exercise, you will:\n",
    "\n",
    "* Extract faces from the images using OpenCV for face detection, and use the same library to apply preprocessing steps;\n",
    "* Run HFS for classification.\n",
    "\n",
    "### Implementation\n",
    "\n",
    "Choose the hyperparameters and run HFS for face recognition, using both the small and large dataset. You can try to change the preprocessing steps (e.g. equalizeHist, GaussianBlur) applied to the images.\n",
    "\n",
    "**Important**: make sure your HFS code is able to handle more than two classes!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Question 2.1 - How did you manage to label more than two classes?\n",
    "\n",
    "To manage more than $2$ classes, we can use one-hot encoding. Instead of having scalar values in $\\{-1,+1\\}$ for labeled data and the solution as a vector of size ```n_samples```, a label is represented by a vector that is equals to $0$ everywhere but when the index equals the label (for instance, if there are 3 classes, a point of class $2$ would have an assigned label of $[0,1,0]$). The resulting solution is a vector of shape (```num_samples```,```num_classes```). Unlabeled data are associated with the label $0$, being the vector of size ```num_classes``` full of $0$.\n",
    "\n",
    "### Question 2.2 - Report the best accuracy you obtained for both (small and augmented) datasets.\n",
    "\n",
    "* Tips:\n",
    "    * The small dataset (10 images per person) is loaded with `load_image_data`.\n",
    "    * Use `load_image_data_augmented` for the augmented dataset (50 images per person). \n",
    "    \n",
    "First, several preprocessing steps can be tried out. For instance one can try to change how the blur is applied, try out different filter, etc. I tried several changes that, for the main part, had no influence (e.g. MedianBlur, other normalization by dividing by the variance). One change that had a major influence was to use another histogram equalization technique : Contrast Limited Adaptive Histogram Equalization (CLAHE) implemented on OpenCV. It improved the results from approximately 10% on the augmented dataset.\n",
    "\n",
    "I used Optuna for the tuning of the other parameters, like $k$, laplacian regularization, c_u, etc. Optuna is a library that automates the parameters' search process.\n",
    "\n",
    "The obtained accuracies are the following:\n",
    "\n",
    "\n",
    "| Dataset | Small | Augmented| \n",
    "|------|------|------|\n",
    "|   Accuracy  | 0.85 | 0.716 |\n",
    "\n",
    "### Question 2.3 - If the accuracy changes when using the augmented dataset, explain why. Does using additional data always increase the performance?\n",
    "\n",
    "The accuracy almost always decreases with the augmented dataset, which can seem counter-intuitive, since the algorithm has more data at its disposal to cluster with. \n",
    "\n",
    "In the 2-moon-dataset case, the high number of points allowed the algorithm to find neighborhood and structure, and therefore it allowed the algorithm to cluster efficiently the datapoints. \n",
    "\n",
    "In the example of the faces, it seems that this structure is less obvious (maybe because the $50$ images is not enough to cover all the space and to create accurate clusters). Therefore, what reveals to be more important in this case is not the total number of data and their underlying structure, but the proportion of data that is labeled. In the augmented dataset there are more images per class ($50$) but the number of labeled data is the same ($4$), so $8$% of the images are already well-classified, while this number is up to $40$% in the small dataset. This simple fact explains the difference of accuracy between the $2$ datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "from imageio import imread\n",
    "import numpy as np\n",
    "import cv2\n",
    "import os\n",
    "\n",
    "from load_images import load_image_data, plot_image_data\n",
    "from load_images import load_image_data_augmented, plot_image_data_augmented"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Function to preprocess the images\n",
    "# You may try to change it and check the impact on the classification accuracy\n",
    "def preprocess_image(image):\n",
    "    \"\"\"\n",
    "    Parameters\n",
    "    ----------\n",
    "    image : array\n",
    "        (width, height) array representing a grayscale image\n",
    "    \n",
    "    Returns\n",
    "    -------\n",
    "        (96, 96) preprocessed image\n",
    "    \"\"\"\n",
    "    output_frame_size = 96   # do not change the output frame size!\n",
    "    image = cv2.bilateralFilter(image, 9, 75, 75)\n",
    "    clahe = cv2.createCLAHE()\n",
    "    image = clahe.apply(image)\n",
    "    image = cv2.GaussianBlur(image, (5, 5), 0)\n",
    "    im = cv2.resize(image, (output_frame_size, output_frame_size)).astype(np.float)\n",
    "    im -= im.mean()\n",
    "    im /= im.max()\n",
    "    image = im\n",
    "    return image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 344,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x720 with 500 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# # 10 images per person\n",
    "np.random.seed(456)   # set seed, since labels are masked randomly\n",
    "images, labels, masked_labels = load_image_data(preprocess_image)\n",
    "\n",
    "# 50 images per person\n",
    "images_a, labels_a, masked_labels_a = load_image_data_augmented(preprocess_image)\n",
    "plot_image_data_augmented(images_a)\n",
    "\n",
    "# Uncomment below if you want to visualize the images\n",
    "# plot_image_data(images)\n",
    "# print(images.shape)\n",
    "# print(masked_labels.reshape(-1, 10))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "!pip install optuna"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 265,
   "metadata": {},
   "outputs": [],
   "source": [
    "import optuna"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 345,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\u001b[32m[I 2021-03-01 18:39:02,857]\u001b[0m A new study created in memory with name: no-name-e4bb2b51-129b-41cb-8424-e98a9523b8a8\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:39:06,507]\u001b[0m Trial 0 finished with value: -0.704 and parameters: {'laplacian_regularization': 0.795288752816869, 'var': 515.7174329447507, 'k': 11, 'laplacian_normalization': 'sym', 'c_u': 0.3352887559429627}. Best is trial 0 with value: -0.704.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:39:10,183]\u001b[0m Trial 1 finished with value: -0.7 and parameters: {'laplacian_regularization': 0.9507760700282751, 'var': 217.912508945438, 'k': 15, 'laplacian_normalization': 'unn', 'c_u': 0.21343868893960538}. Best is trial 0 with value: -0.704.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:39:13,592]\u001b[0m Trial 2 finished with value: -0.65 and parameters: {'laplacian_regularization': 0.4238731599852732, 'var': 426.43535136612275, 'k': 39, 'laplacian_normalization': 'rw', 'c_u': 0.36400720885483967}. Best is trial 0 with value: -0.704.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:39:17,146]\u001b[0m Trial 3 finished with value: -0.348 and parameters: {'laplacian_regularization': 0.26160434290651924, 'var': 588.7210889393303, 'k': 2, 'laplacian_normalization': 'sym', 'c_u': 0.1708406492572051}. Best is trial 0 with value: -0.704.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:39:20,758]\u001b[0m Trial 4 finished with value: -0.632 and parameters: {'laplacian_regularization': 0.6698343017559388, 'var': 954.8544048194484, 'k': 31, 'laplacian_normalization': 'unn', 'c_u': 0.055100802020059514}. Best is trial 0 with value: -0.704.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:39:24,660]\u001b[0m Trial 5 finished with value: -0.694 and parameters: {'laplacian_regularization': 0.3222604423880081, 'var': 840.14797962482, 'k': 13, 'laplacian_normalization': 'sym', 'c_u': 0.3659774909415063}. Best is trial 0 with value: -0.704.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:39:28,682]\u001b[0m Trial 6 finished with value: -0.686 and parameters: {'laplacian_regularization': 0.2579361486724271, 'var': 114.53404196305718, 'k': 34, 'laplacian_normalization': 'rw', 'c_u': 0.14173002675634044}. Best is trial 0 with value: -0.704.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:39:32,267]\u001b[0m Trial 7 finished with value: -0.698 and parameters: {'laplacian_regularization': 0.03194450124134851, 'var': 975.1252659799532, 'k': 15, 'laplacian_normalization': 'sym', 'c_u': 0.4215074066120743}. Best is trial 0 with value: -0.704.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:39:36,010]\u001b[0m Trial 8 finished with value: -0.706 and parameters: {'laplacian_regularization': 0.36118255041294156, 'var': 64.06049381599178, 'k': 23, 'laplacian_normalization': 'sym', 'c_u': 0.32134300823384004}. Best is trial 8 with value: -0.706.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:39:39,566]\u001b[0m Trial 9 finished with value: -0.674 and parameters: {'laplacian_regularization': 0.8646348647988683, 'var': 974.0776622378966, 'k': 21, 'laplacian_normalization': 'unn', 'c_u': 0.15444393856652777}. Best is trial 8 with value: -0.706.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:39:43,382]\u001b[0m Trial 10 finished with value: -0.706 and parameters: {'laplacian_regularization': 0.5941357391751795, 'var': 10.679244337928587, 'k': 23, 'laplacian_normalization': 'sym', 'c_u': 0.28489123885433554}. Best is trial 8 with value: -0.706.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:39:46,966]\u001b[0m Trial 11 finished with value: -0.704 and parameters: {'laplacian_regularization': 0.5729624044242914, 'var': 71.62459798770196, 'k': 25, 'laplacian_normalization': 'sym', 'c_u': 0.28051675877438337}. Best is trial 8 with value: -0.706.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:39:50,527]\u001b[0m Trial 12 finished with value: -0.704 and parameters: {'laplacian_regularization': 0.047679577805793794, 'var': 10.760466659904495, 'k': 25, 'laplacian_normalization': 'sym', 'c_u': 0.4722105912075357}. Best is trial 8 with value: -0.706.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:39:54,304]\u001b[0m Trial 13 finished with value: -0.686 and parameters: {'laplacian_regularization': 0.5275943846545731, 'var': 252.88534415037014, 'k': 30, 'laplacian_normalization': 'sym', 'c_u': 0.2802862532057285}. Best is trial 8 with value: -0.706.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:39:58,231]\u001b[0m Trial 14 finished with value: -0.696 and parameters: {'laplacian_regularization': 0.6532496809018857, 'var': 314.1308345034809, 'k': 21, 'laplacian_normalization': 'sym', 'c_u': 0.26180698012613857}. Best is trial 8 with value: -0.706.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:40:01,907]\u001b[0m Trial 15 finished with value: -0.694 and parameters: {'laplacian_regularization': 0.40389301389731047, 'var': 19.33914254844776, 'k': 6, 'laplacian_normalization': 'rw', 'c_u': 0.4947297984216177}. Best is trial 8 with value: -0.706.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:40:05,588]\u001b[0m Trial 16 finished with value: -0.67 and parameters: {'laplacian_regularization': 0.15253515874840307, 'var': 714.7524098952888, 'k': 26, 'laplacian_normalization': 'sym', 'c_u': 0.4213227030249475}. Best is trial 8 with value: -0.706.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:40:09,490]\u001b[0m Trial 17 finished with value: -0.682 and parameters: {'laplacian_regularization': 0.7309639298407642, 'var': 153.95967153873596, 'k': 40, 'laplacian_normalization': 'sym', 'c_u': 0.3276797924616881}. Best is trial 8 with value: -0.706.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:40:13,118]\u001b[0m Trial 18 finished with value: -0.69 and parameters: {'laplacian_regularization': 0.462216090600301, 'var': 367.07414714359203, 'k': 18, 'laplacian_normalization': 'sym', 'c_u': 0.018729283514640704}. Best is trial 8 with value: -0.706.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:40:16,778]\u001b[0m Trial 19 finished with value: -0.67 and parameters: {'laplacian_regularization': 0.5772752986992872, 'var': 3.402244006627093, 'k': 34, 'laplacian_normalization': 'rw', 'c_u': 0.2157513163400754}. Best is trial 8 with value: -0.706.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:40:20,185]\u001b[0m Trial 20 finished with value: -0.702 and parameters: {'laplacian_regularization': 0.33527689608965733, 'var': 148.12517462385867, 'k': 19, 'laplacian_normalization': 'unn', 'c_u': 0.43651551313477244}. Best is trial 8 with value: -0.706.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:40:23,770]\u001b[0m Trial 21 finished with value: -0.704 and parameters: {'laplacian_regularization': 0.015165369852458699, 'var': 29.33153952484821, 'k': 25, 'laplacian_normalization': 'sym', 'c_u': 0.4453488276578861}. Best is trial 8 with value: -0.706.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:40:29,160]\u001b[0m Trial 22 finished with value: -0.686 and parameters: {'laplacian_regularization': 0.1111683029719232, 'var': 21.969067730633306, 'k': 27, 'laplacian_normalization': 'sym', 'c_u': 0.4721499846939124}. Best is trial 8 with value: -0.706.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:40:33,788]\u001b[0m Trial 23 finished with value: -0.7 and parameters: {'laplacian_regularization': 0.8176292755255843, 'var': 549.862653770764, 'k': 11, 'laplacian_normalization': 'sym', 'c_u': 0.340670356393261}. Best is trial 8 with value: -0.706.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:40:37,396]\u001b[0m Trial 24 finished with value: -0.702 and parameters: {'laplacian_regularization': 0.7684885776094716, 'var': 661.3542376682976, 'k': 10, 'laplacian_normalization': 'sym', 'c_u': 0.3134176152857065}. Best is trial 8 with value: -0.706.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:40:40,923]\u001b[0m Trial 25 finished with value: -0.682 and parameters: {'laplacian_regularization': 0.6471018151567439, 'var': 505.6007316153117, 'k': 7, 'laplacian_normalization': 'sym', 'c_u': 0.38702444951384063}. Best is trial 8 with value: -0.706.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:40:44,438]\u001b[0m Trial 26 finished with value: -0.698 and parameters: {'laplacian_regularization': 0.9728712195827147, 'var': 424.7970277653377, 'k': 22, 'laplacian_normalization': 'sym', 'c_u': 0.21817723751060247}. Best is trial 8 with value: -0.706.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:40:47,935]\u001b[0m Trial 27 finished with value: -0.686 and parameters: {'laplacian_regularization': 0.8615203252030226, 'var': 768.1594827390525, 'k': 16, 'laplacian_normalization': 'sym', 'c_u': 0.3141932403845407}. Best is trial 8 with value: -0.706.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:40:51,410]\u001b[0m Trial 28 finished with value: -0.712 and parameters: {'laplacian_regularization': 0.11489409493235105, 'var': 212.4495080763357, 'k': 23, 'laplacian_normalization': 'sym', 'c_u': 0.097209757851769}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:40:55,020]\u001b[0m Trial 29 finished with value: -0.688 and parameters: {'laplacian_regularization': 0.16844866376960574, 'var': 220.0149326523213, 'k': 29, 'laplacian_normalization': 'sym', 'c_u': 0.11178158290050044}. Best is trial 28 with value: -0.712.\u001b[0m\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\u001b[32m[I 2021-03-01 18:40:59,083]\u001b[0m Trial 30 finished with value: -0.704 and parameters: {'laplacian_regularization': 0.22633623580150092, 'var': 99.01864190781237, 'k': 23, 'laplacian_normalization': 'sym', 'c_u': 0.05120568283144048}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:41:02,913]\u001b[0m Trial 31 finished with value: -0.694 and parameters: {'laplacian_regularization': 0.360236369436459, 'var': 309.95260218738565, 'k': 17, 'laplacian_normalization': 'sym', 'c_u': 0.23913951704043157}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:41:06,714]\u001b[0m Trial 32 finished with value: -0.698 and parameters: {'laplacian_regularization': 0.5182492311963082, 'var': 433.54489936411994, 'k': 19, 'laplacian_normalization': 'unn', 'c_u': 0.29101488339790926}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:41:10,225]\u001b[0m Trial 33 finished with value: -0.686 and parameters: {'laplacian_regularization': 0.9104543907067997, 'var': 617.0071002293829, 'k': 13, 'laplacian_normalization': 'sym', 'c_u': 0.3598820475500726}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:41:13,983]\u001b[0m Trial 34 finished with value: -0.708 and parameters: {'laplacian_regularization': 0.07800918751540645, 'var': 159.30075846591268, 'k': 24, 'laplacian_normalization': 'sym', 'c_u': 0.19283990598988265}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:41:17,522]\u001b[0m Trial 35 finished with value: -0.688 and parameters: {'laplacian_regularization': 0.07571965369722333, 'var': 187.5857252996266, 'k': 28, 'laplacian_normalization': 'sym', 'c_u': 0.0829465216922872}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:41:22,868]\u001b[0m Trial 36 finished with value: -0.692 and parameters: {'laplacian_regularization': 0.1998814551433382, 'var': 71.6275548031303, 'k': 32, 'laplacian_normalization': 'rw', 'c_u': 0.17990921575840102}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:41:27,560]\u001b[0m Trial 37 finished with value: -0.71 and parameters: {'laplacian_regularization': 0.29058078769263185, 'var': 285.27642076881943, 'k': 23, 'laplacian_normalization': 'sym', 'c_u': 0.11732617955848411}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:41:32,055]\u001b[0m Trial 38 finished with value: -0.682 and parameters: {'laplacian_regularization': 0.28301511472109825, 'var': 252.6185981224006, 'k': 37, 'laplacian_normalization': 'sym', 'c_u': 0.11258866524495721}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:41:36,524]\u001b[0m Trial 39 finished with value: -0.7 and parameters: {'laplacian_regularization': 0.11514718056543671, 'var': 316.343986496501, 'k': 23, 'laplacian_normalization': 'unn', 'c_u': 0.17494537635658036}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:41:40,571]\u001b[0m Trial 40 finished with value: -0.69 and parameters: {'laplacian_regularization': 0.384093399178953, 'var': 191.2922904524178, 'k': 32, 'laplacian_normalization': 'sym', 'c_u': 0.13280743392688513}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:41:44,120]\u001b[0m Trial 41 finished with value: -0.704 and parameters: {'laplacian_regularization': 0.45645154828481815, 'var': 128.10748874632878, 'k': 24, 'laplacian_normalization': 'sym', 'c_u': 0.06763261004446347}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:41:48,583]\u001b[0m Trial 42 finished with value: -0.702 and parameters: {'laplacian_regularization': 0.30492927096447936, 'var': 73.77799137881424, 'k': 20, 'laplacian_normalization': 'sym', 'c_u': 0.18829627410312452}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:41:52,409]\u001b[0m Trial 43 finished with value: -0.69 and parameters: {'laplacian_regularization': 0.23772184673881025, 'var': 262.749313559868, 'k': 27, 'laplacian_normalization': 'sym', 'c_u': 0.24224620853731868}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:41:55,829]\u001b[0m Trial 44 finished with value: -0.706 and parameters: {'laplacian_regularization': 0.5780774834521911, 'var': 180.80919821859956, 'k': 22, 'laplacian_normalization': 'sym', 'c_u': 0.102672776312871}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:41:59,974]\u001b[0m Trial 45 finished with value: -0.71 and parameters: {'laplacian_regularization': 0.0004971058876589857, 'var': 200.1096509310455, 'k': 21, 'laplacian_normalization': 'sym', 'c_u': 0.004337779790659968}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:42:03,358]\u001b[0m Trial 46 finished with value: -0.688 and parameters: {'laplacian_regularization': 0.06958682154205219, 'var': 389.55254770061885, 'k': 21, 'laplacian_normalization': 'sym', 'c_u': 0.013135945668560162}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:42:06,888]\u001b[0m Trial 47 finished with value: -0.704 and parameters: {'laplacian_regularization': 0.128751268828772, 'var': 280.07025204032686, 'k': 14, 'laplacian_normalization': 'rw', 'c_u': 0.03383214058796488}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:42:09,574]\u001b[0m Trial 48 finished with value: -0.706 and parameters: {'laplacian_regularization': 0.18749801989990128, 'var': 198.87604058494637, 'k': 17, 'laplacian_normalization': 'sym', 'c_u': 0.09395942899739312}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:42:13,938]\u001b[0m Trial 49 finished with value: -0.698 and parameters: {'laplacian_regularization': 0.006230795964772107, 'var': 361.47465928994455, 'k': 16, 'laplacian_normalization': 'sym', 'c_u': 0.1482169242166847}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:42:17,626]\u001b[0m Trial 50 finished with value: -0.704 and parameters: {'laplacian_regularization': 0.0023257815379519875, 'var': 225.30582219736027, 'k': 25, 'laplacian_normalization': 'sym', 'c_u': 0.04526523866073926}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:42:21,748]\u001b[0m Trial 51 finished with value: -0.704 and parameters: {'laplacian_regularization': 0.19119075435137406, 'var': 178.9240176193745, 'k': 19, 'laplacian_normalization': 'sym', 'c_u': 0.08612835724996605}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:42:25,475]\u001b[0m Trial 52 finished with value: -0.706 and parameters: {'laplacian_regularization': 0.08239752631308056, 'var': 152.997941776448, 'k': 21, 'laplacian_normalization': 'sym', 'c_u': 0.12023991310258181}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:42:29,076]\u001b[0m Trial 53 finished with value: -0.704 and parameters: {'laplacian_regularization': 0.07676822704468705, 'var': 107.41255673877234, 'k': 17, 'laplacian_normalization': 'sym', 'c_u': 0.12832928887261397}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:42:32,620]\u001b[0m Trial 54 finished with value: -0.71 and parameters: {'laplacian_regularization': 0.15387026005730126, 'var': 207.6528344704787, 'k': 22, 'laplacian_normalization': 'sym', 'c_u': 0.09072749051118201}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:42:36,536]\u001b[0m Trial 55 finished with value: -0.682 and parameters: {'laplacian_regularization': 0.03540328197833452, 'var': 291.8117220277274, 'k': 24, 'laplacian_normalization': 'sym', 'c_u': 0.00039941438869306456}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:42:41,741]\u001b[0m Trial 56 finished with value: -0.688 and parameters: {'laplacian_regularization': 0.1413874137017847, 'var': 240.95730838875699, 'k': 27, 'laplacian_normalization': 'unn', 'c_u': 0.19491665815023607}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:42:47,631]\u001b[0m Trial 57 finished with value: -0.702 and parameters: {'laplacian_regularization': 0.7067896122599051, 'var': 158.3961462233614, 'k': 22, 'laplacian_normalization': 'sym', 'c_u': 0.15437395180413288}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:42:51,054]\u001b[0m Trial 58 finished with value: -0.704 and parameters: {'laplacian_regularization': 0.6274853040718993, 'var': 219.44338643712194, 'k': 20, 'laplacian_normalization': 'sym', 'c_u': 0.06723818027664086}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:42:54,128]\u001b[0m Trial 59 finished with value: -0.702 and parameters: {'laplacian_regularization': 0.28118240865042027, 'var': 47.20809507527122, 'k': 26, 'laplacian_normalization': 'rw', 'c_u': 0.026424519677656275}. Best is trial 28 with value: -0.712.\u001b[0m\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\u001b[32m[I 2021-03-01 18:42:57,535]\u001b[0m Trial 60 finished with value: -0.682 and parameters: {'laplacian_regularization': 0.09468799585986203, 'var': 134.44512447453153, 'k': 29, 'laplacian_normalization': 'sym', 'c_u': 0.1272450903420625}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:43:01,330]\u001b[0m Trial 61 finished with value: -0.704 and parameters: {'laplacian_regularization': 0.05104600600175716, 'var': 105.52151137498525, 'k': 24, 'laplacian_normalization': 'sym', 'c_u': 0.20818304263754803}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:43:05,228]\u001b[0m Trial 62 finished with value: -0.704 and parameters: {'laplacian_regularization': 0.21972640002275423, 'var': 62.44494648091748, 'k': 21, 'laplacian_normalization': 'sym', 'c_u': 0.16494420910393814}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:43:08,866]\u001b[0m Trial 63 finished with value: -0.706 and parameters: {'laplacian_regularization': 0.1687977081661038, 'var': 338.87289169413674, 'k': 19, 'laplacian_normalization': 'sym', 'c_u': 0.09522439378527517}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:43:12,505]\u001b[0m Trial 64 finished with value: -0.694 and parameters: {'laplacian_regularization': 0.5485620941751168, 'var': 303.5505839975781, 'k': 22, 'laplacian_normalization': 'sym', 'c_u': 0.07005429769078313}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:43:16,036]\u001b[0m Trial 65 finished with value: -0.704 and parameters: {'laplacian_regularization': 0.15810292216226224, 'var': 348.56164099948677, 'k': 18, 'laplacian_normalization': 'sym', 'c_u': 0.09871476735105227}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:43:19,600]\u001b[0m Trial 66 finished with value: -0.706 and parameters: {'laplacian_regularization': 0.2543867380618074, 'var': 209.37843879991271, 'k': 26, 'laplacian_normalization': 'sym', 'c_u': 0.10375140925265437}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:43:23,151]\u001b[0m Trial 67 finished with value: -0.708 and parameters: {'laplacian_regularization': 0.20436283562050683, 'var': 175.5701188237617, 'k': 23, 'laplacian_normalization': 'sym', 'c_u': 0.14408756310380896}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:43:26,702]\u001b[0m Trial 68 finished with value: -0.706 and parameters: {'laplacian_regularization': 0.2388970240665727, 'var': 269.21929337976945, 'k': 26, 'laplacian_normalization': 'sym', 'c_u': 0.1410933700933668}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:43:30,256]\u001b[0m Trial 69 finished with value: -0.688 and parameters: {'laplacian_regularization': 0.33632962675411254, 'var': 217.38648714822133, 'k': 29, 'laplacian_normalization': 'sym', 'c_u': 0.16074320042872015}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:43:34,148]\u001b[0m Trial 70 finished with value: -0.706 and parameters: {'laplacian_regularization': 0.10706110057805104, 'var': 276.1420595145983, 'k': 25, 'laplacian_normalization': 'sym', 'c_u': 0.14707927482797084}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:43:37,723]\u001b[0m Trial 71 finished with value: -0.71 and parameters: {'laplacian_regularization': 0.1928002172866683, 'var': 183.54100526535265, 'k': 23, 'laplacian_normalization': 'sym', 'c_u': 0.08270727656636703}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:43:41,432]\u001b[0m Trial 72 finished with value: -0.702 and parameters: {'laplacian_regularization': 0.16503650633601136, 'var': 473.38166465189977, 'k': 23, 'laplacian_normalization': 'sym', 'c_u': 0.07668856900256696}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:43:45,366]\u001b[0m Trial 73 finished with value: -0.706 and parameters: {'laplacian_regularization': 0.19712798617893332, 'var': 341.2221444561175, 'k': 24, 'laplacian_normalization': 'sym', 'c_u': 0.05256044781175337}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:43:48,906]\u001b[0m Trial 74 finished with value: -0.708 and parameters: {'laplacian_regularization': 0.13156462200720553, 'var': 182.597032433631, 'k': 22, 'laplacian_normalization': 'sym', 'c_u': 0.11369730906940426}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:43:52,579]\u001b[0m Trial 75 finished with value: -0.708 and parameters: {'laplacian_regularization': 0.12640577152691845, 'var': 146.10531663313068, 'k': 20, 'laplacian_normalization': 'sym', 'c_u': 0.11342319810227483}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:43:56,606]\u001b[0m Trial 76 finished with value: -0.708 and parameters: {'laplacian_regularization': 0.03191702840623353, 'var': 175.09587299025668, 'k': 22, 'laplacian_normalization': 'unn', 'c_u': 0.1368251363022542}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:44:01,414]\u001b[0m Trial 77 finished with value: -0.706 and parameters: {'laplacian_regularization': 0.03204588174579554, 'var': 232.4235030116896, 'k': 22, 'laplacian_normalization': 'unn', 'c_u': 0.13642899805697273}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:44:05,463]\u001b[0m Trial 78 finished with value: -0.708 and parameters: {'laplacian_regularization': 0.13611701123928588, 'var': 129.6985161425713, 'k': 20, 'laplacian_normalization': 'sym', 'c_u': 0.0011360115405327446}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:44:09,039]\u001b[0m Trial 79 finished with value: -0.702 and parameters: {'laplacian_regularization': 0.2167528560972054, 'var': 101.34516144213542, 'k': 18, 'laplacian_normalization': 'sym', 'c_u': 0.062412409444631126}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:44:12,775]\u001b[0m Trial 80 finished with value: -0.704 and parameters: {'laplacian_regularization': 0.2641707849576596, 'var': 121.34097240766714, 'k': 24, 'laplacian_normalization': 'sym', 'c_u': 0.0031014903042913167}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:44:16,504]\u001b[0m Trial 81 finished with value: -0.712 and parameters: {'laplacian_regularization': 0.06348137698167824, 'var': 147.83885127078048, 'k': 20, 'laplacian_normalization': 'unn', 'c_u': 0.041609739970567626}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:44:20,215]\u001b[0m Trial 82 finished with value: -0.708 and parameters: {'laplacian_regularization': 0.1099843876522328, 'var': 150.8846251885298, 'k': 20, 'laplacian_normalization': 'unn', 'c_u': 0.03832042115695551}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:44:24,116]\u001b[0m Trial 83 finished with value: -0.712 and parameters: {'laplacian_regularization': 0.06884002568757952, 'var': 164.46211535936706, 'k': 23, 'laplacian_normalization': 'unn', 'c_u': 0.0488052490362789}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:44:27,995]\u001b[0m Trial 84 finished with value: -0.704 and parameters: {'laplacian_regularization': 0.05545056653207018, 'var': 247.3197439408454, 'k': 23, 'laplacian_normalization': 'unn', 'c_u': 0.046286844936685174}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:44:31,770]\u001b[0m Trial 85 finished with value: -0.702 and parameters: {'laplacian_regularization': 0.017870839052247955, 'var': 77.61517686614195, 'k': 18, 'laplacian_normalization': 'unn', 'c_u': 0.01746373764916298}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:44:35,626]\u001b[0m Trial 86 finished with value: -0.708 and parameters: {'laplacian_regularization': 0.05694730230607362, 'var': 198.45365458568344, 'k': 21, 'laplacian_normalization': 'unn', 'c_u': 0.038430943615046606}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:44:39,161]\u001b[0m Trial 87 finished with value: -0.69 and parameters: {'laplacian_regularization': 0.09203329313186277, 'var': 168.2609871182657, 'k': 27, 'laplacian_normalization': 'unn', 'c_u': 0.08455218936340768}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:44:42,731]\u001b[0m Trial 88 finished with value: -0.71 and parameters: {'laplacian_regularization': 0.14028338374519822, 'var': 128.5866767461912, 'k': 20, 'laplacian_normalization': 'unn', 'c_u': 0.026209358676420803}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:44:46,403]\u001b[0m Trial 89 finished with value: -0.71 and parameters: {'laplacian_regularization': 3.145015307440269e-05, 'var': 88.44067371006548, 'k': 16, 'laplacian_normalization': 'unn', 'c_u': 0.03137818429126428}. Best is trial 28 with value: -0.712.\u001b[0m\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\u001b[32m[I 2021-03-01 18:44:50,047]\u001b[0m Trial 90 finished with value: -0.702 and parameters: {'laplacian_regularization': 0.0015902818754411213, 'var': 47.95438280758292, 'k': 14, 'laplacian_normalization': 'unn', 'c_u': 0.017458778805136757}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:44:53,853]\u001b[0m Trial 91 finished with value: -0.704 and parameters: {'laplacian_regularization': 0.08872740157496234, 'var': 96.90414252118923, 'k': 19, 'laplacian_normalization': 'unn', 'c_u': 0.05987747201896628}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:44:57,208]\u001b[0m Trial 92 finished with value: -0.71 and parameters: {'laplacian_regularization': 0.17171052051418467, 'var': 201.9071340692897, 'k': 16, 'laplacian_normalization': 'unn', 'c_u': 0.07746869115718742}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:45:00,627]\u001b[0m Trial 93 finished with value: -0.71 and parameters: {'laplacian_regularization': 0.17449063755536537, 'var': 251.29790329738293, 'k': 11, 'laplacian_normalization': 'unn', 'c_u': 0.03417493365229303}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:45:04,032]\u001b[0m Trial 94 finished with value: -0.706 and parameters: {'laplacian_regularization': 0.17249174040238316, 'var': 241.26483494848245, 'k': 12, 'laplacian_normalization': 'unn', 'c_u': 0.03338762973974575}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:45:07,637]\u001b[0m Trial 95 finished with value: -0.71 and parameters: {'laplacian_regularization': 0.1844210594505017, 'var': 83.72898094934052, 'k': 10, 'laplacian_normalization': 'unn', 'c_u': 0.027487876917901777}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:45:11,203]\u001b[0m Trial 96 finished with value: -0.696 and parameters: {'laplacian_regularization': 0.1451835955713041, 'var': 30.774330963157126, 'k': 7, 'laplacian_normalization': 'unn', 'c_u': 0.0774049647114511}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:45:14,595]\u001b[0m Trial 97 finished with value: -0.67 and parameters: {'laplacian_regularization': 0.30268425167692903, 'var': 205.83459434926837, 'k': 5, 'laplacian_normalization': 'unn', 'c_u': 0.0552849028565028}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:45:18,262]\u001b[0m Trial 98 finished with value: -0.71 and parameters: {'laplacian_regularization': 0.11366251308089369, 'var': 83.78667622867013, 'k': 10, 'laplacian_normalization': 'unn', 'c_u': 0.009030241684568703}. Best is trial 28 with value: -0.712.\u001b[0m\n",
      "\u001b[32m[I 2021-03-01 18:45:22,013]\u001b[0m Trial 99 finished with value: -0.716 and parameters: {'laplacian_regularization': 0.11080189362149373, 'var': 125.05540592439401, 'k': 9, 'laplacian_normalization': 'unn', 'c_u': 0.004888044132053262}. Best is trial 99 with value: -0.716.\u001b[0m\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'laplacian_regularization': 0.11080189362149373,\n",
       " 'var': 125.05540592439401,\n",
       " 'k': 9,\n",
       " 'laplacian_normalization': 'unn',\n",
       " 'c_u': 0.004888044132053262}"
      ]
     },
     "execution_count": 345,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def objective(trial):\n",
    "    params_face_rec['laplacian_regularization'] = trial.suggest_float(\"laplacian_regularization\", 0, 1)\n",
    "    params_face_rec['var'] = trial.suggest_float(\"var\", 1, 1000)\n",
    "    params_face_rec['eps'] = 0\n",
    "    params_face_rec['k'] = trial.suggest_int(\"k\", 1, 40)\n",
    "    params_face_rec['laplacian_normalization'] = trial.suggest_categorical(\"laplacian_normalization\", [\"unn\", \"rw\", \"sym\"])\n",
    "    params_face_rec['c_l'] = 1\n",
    "    params_face_rec['c_u'] = trial.suggest_float(\"c_u\", 0.0, 0.5)\n",
    "    \n",
    "    # graph Laplacian\n",
    "    L = build_laplacian_regularized(images_a, \n",
    "                                params_face_rec['laplacian_regularization'], \n",
    "                                params_face_rec['var'], \n",
    "                                params_face_rec['eps'], \n",
    "                                params_face_rec['k'], \n",
    "                                params_face_rec['laplacian_normalization'])\n",
    "\n",
    "    # Run HFS\n",
    "    predicted_labels, f = compute_hfs(L, masked_labels_a, soft=True, **params_face_rec)\n",
    "    accuracy = np.equal(predicted_labels, labels_a).mean()\n",
    "    \n",
    "    return - accuracy\n",
    "\n",
    "study = optuna.create_study()\n",
    "study.optimize(objective, n_trials=100)\n",
    "study.best_params"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 380,
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "Define parameters for face recognition with HFS\n",
    "\"\"\"\n",
    "params_face_rec = {}\n",
    "params_face_rec['laplacian_regularization'] = 0.1\n",
    "params_face_rec['var'] = 125.0\n",
    "params_face_rec['eps'] = None\n",
    "params_face_rec['k'] = 9\n",
    "params_face_rec['laplacian_normalization'] = 'unn'\n",
    "params_face_rec['c_l'] = 1\n",
    "params_face_rec['c_u'] = 0.2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 381,
   "metadata": {},
   "outputs": [],
   "source": [
    "# graph Laplacian\n",
    "L = build_laplacian_regularized(images, \n",
    "                                params_face_rec['laplacian_regularization'], \n",
    "                                params_face_rec['var'], \n",
    "                                params_face_rec['eps'], \n",
    "                                params_face_rec['k'], \n",
    "                                params_face_rec['laplacian_normalization'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 382,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy =  0.85\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Run HFS\n",
    "predicted_labels, f = compute_hfs(L, masked_labels, soft=True, **params_face_rec)\n",
    "accuracy = np.equal(predicted_labels, labels).mean()\n",
    "print(\"Accuracy = \", accuracy)\n",
    "\n",
    "# Visualize predicted vs true labels\n",
    "plt.subplot(121)\n",
    "plt.imshow(labels.reshape((-1, 10)))\n",
    "plt.subplot(122)\n",
    "plt.imshow(predicted_labels.reshape((-1, 10)))\n",
    "plt.title(\"Accuracy: {}\".format(accuracy))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 383,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy =  0.716\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# graph Laplacian\n",
    "L = build_laplacian_regularized(images_a, \n",
    "                                params_face_rec['laplacian_regularization'], \n",
    "                                params_face_rec['var'], \n",
    "                                params_face_rec['eps'], \n",
    "                                params_face_rec['k'], \n",
    "                                params_face_rec['laplacian_normalization'])\n",
    "# Run HFS\n",
    "predicted_labels, f = compute_hfs(L, masked_labels_a, soft=True, **params_face_rec)\n",
    "accuracy = np.equal(predicted_labels, labels_a).mean()\n",
    "print(\"Accuracy = \", accuracy)\n",
    "\n",
    "# Visualize predicted vs true labels\n",
    "plt.subplot(121)\n",
    "plt.imshow(labels.reshape((-1, 10)))\n",
    "plt.subplot(122)\n",
    "plt.imshow(predicted_labels.reshape((-1, 10)))\n",
    "plt.title(\"Accuracy: {}\".format(accuracy))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. Online SSL\n",
    "\n",
    "Now, instead of having all the data available at once, images will be received online: at each time $t$, a new image $x_t$ is observed and the algorithm has to output a label $y_t$. \n",
    "\n",
    "Use the function `create_user_profile` to capture a training set of labeled data (of your face and someone else). The faces will be preprocessed and saved in the folder `data/faces`. They will be loaded by `online_face_recognition`.\n",
    "\n",
    "\n",
    "### Implementation\n",
    "\n",
    "Choose the hyperparameters and complete the functions `online_ssl_update_centroids` and `online_ssl_compute_solution`. \n",
    "\n",
    "Modify your code to be able to disregard faces it cannot recognize.\n",
    "\n",
    "* Tips:\n",
    "    * You can use the functions `build_similarity_graph` and `build_laplacian`."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Question 3.1 - Attach to this notebook some of the resulting frames of online face recognition. \n",
    "\n",
    "* Tips: \n",
    "    * You can save the resulting frame and add it to the notebook in a markdown cell as `![title](picture.png)`\n",
    "\n",
    "I used the parameters from the previous part. The overall results are quite good, but since Inès and I look similar, we had some trouble calibrating the algorithm and saving the pictures when creating the profiles. Once calibrated, the method worked well.   "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![Online recognition of Pierre and Ines](results/frame_p_i.png)\n",
    "![Online recognition of Pierre and Diana](results/frame_p_d.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Question 3.2 - What strategy did you use to label a face as unknown? Attach to this notebook an example of a unknown face being correctly labeled as unknown.\n",
    "\n",
    "* Tips\n",
    "    * If you identify a face as unknown, you can return `[(\"unknown\", score)]` from the function `online_ssl_compute_solution`.\n",
    "\n",
    "To deal with unknown faces, we can note that an unknown face is often correlated to a low score (seems logical!). Therefore, one can put a threshold, fixed here arbitrarily at $0.5$. If none of the label scores are higher than this threshold then the face is qualified as unknown. We can associate a score of $4*(0.5-\\textrm{best_score})$ to the unknow label, that helps identifying how much the face is unknown (the $4$ comes from the fact the best score of an unknown face was around $0.25$, so the factor $4$ made the unknown score be around 1 for unknown faces.)   "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![Online recognition of Pierre and Diana](results/unknown.png)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import cv2 as cv\n",
    "import os\n",
    "import sys\n",
    "from scipy.spatial import distance\n",
    "import scipy.io as sio\n",
    "\n",
    "from helper_online_ssl import create_user_profile, online_face_recognition"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "Define parameters for face recognition with HFS\n",
    "\"\"\"\n",
    "params_online_ssl = {}\n",
    "params_online_ssl['laplacian_regularization'] = 0.1\n",
    "params_online_ssl['var'] = 1000.0\n",
    "params_online_ssl['eps'] = None\n",
    "params_online_ssl['k'] = 20\n",
    "params_online_ssl['laplacian_normalization'] = 'unn'\n",
    "params_online_ssl['c_l'] = 1\n",
    "params_online_ssl['c_u'] = 0.2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Function to preprocess the images\n",
    "# You may try to change it and check the impact on the classification accuracy\n",
    "def preprocess_image(image):\n",
    "    \"\"\"\n",
    "    Parameters\n",
    "    ----------\n",
    "    image : array\n",
    "        (width, height) array representing a grayscale image\n",
    "    \n",
    "    Returns\n",
    "    -------\n",
    "        (96, 96) preprocessed image\n",
    "    \"\"\"\n",
    "    output_frame_size = 96   # do not change the output frame size!\n",
    "    image = cv2.bilateralFilter(image, 9, 75, 75)\n",
    "    image = cv2.equalizeHist(image)\n",
    "    image = cv2.GaussianBlur(image, (5, 5), 0)\n",
    "    im = cv2.resize(image, (output_frame_size, output_frame_size)).astype(np.float)\n",
    "    im -= im.mean()\n",
    "    im /= im.max()\n",
    "    image = im\n",
    "    return image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [],
   "source": [
    "class IncrementalKCenters:\n",
    "    def __init__(self, labeled_faces, labels, label_names, max_num_centroids=50):\n",
    "        #  Number of labels\n",
    "        self.n_labels = max(labels)\n",
    "\n",
    "        #  Dimension of the input image\n",
    "        self.image_dimension = labeled_faces.shape[1]\n",
    "\n",
    "        #  Check input validity\n",
    "        assert (set(labels) == set(\n",
    "            range(1, 1 + self.n_labels))), \"Initially provided faces should be labeled in [1, max]\"\n",
    "        assert (len(labeled_faces) == len(labels)), \"Initial faces and initial labels are not of same size\"\n",
    "\n",
    "        #  Number of labelled faces\n",
    "        self.n_labeled_faces = len(labeled_faces)\n",
    "\n",
    "        # Model parameter : number of maximum stored centroids\n",
    "        self.max_num_centroids = max_num_centroids\n",
    "\n",
    "        # Model centroids (inital labeled faces). Shape = (number_of_centroids, dimension)\n",
    "        self.centroids = labeled_faces\n",
    "\n",
    "        # Centroids labels\n",
    "        self.Y = labels\n",
    "        \n",
    "        # Label names (= user names)\n",
    "        self.label_names = label_names\n",
    "\n",
    "        # Variables that are initialized in online_ssl_update_centroids()\n",
    "        self.centroids_distances = None\n",
    "        self.taboo = None\n",
    "        self.V = None\n",
    "        self.init = True\n",
    "\n",
    "        # index of x_t (initialized later)\n",
    "        self.last_face = None\n",
    "    \n",
    "    def initialize(self):\n",
    "        \"\"\"\n",
    "        Initialization after the first time that the maximum number of centroids is reached.\n",
    "        \"\"\"       \n",
    "        #  Compute the centroids distances\n",
    "        self.centroids_distances = distance.cdist(self.centroids, self.centroids)\n",
    "\n",
    "        #  set labeled nodes and self loops as infinitely distant, to avoid merging labeled centroids\n",
    "        np.fill_diagonal(self.centroids_distances, +np.Inf)\n",
    "        self.centroids_distances[0:self.n_labeled_faces, 0:self.n_labeled_faces] = +np.Inf\n",
    "\n",
    "        # put labeled nodes in the taboo list\n",
    "        self.taboo = np.array(range(self.centroids.shape[0])) < self.n_labeled_faces\n",
    "\n",
    "        # initialize multiplicity\n",
    "        self.V = np.ones(self.centroids.shape[0])\n",
    "\n",
    "\n",
    "    def online_ssl_update_centroids(self, face):\n",
    "        \"\"\"\n",
    "        Update centroids, multiplicity vector V, labels Y.\n",
    "        \n",
    "        Note: In Y, set label to 0 for unlabeled faces.\n",
    "\n",
    "        Parameters\n",
    "        ----------\n",
    "        face : array\n",
    "            New sample\n",
    "        \n",
    "        Returns\n",
    "        --------\n",
    "        List with the scores for each possible label:\n",
    "            [(label_1, score_1), (label_2, score_2), ...]\n",
    "        \"\"\"\n",
    "\n",
    "        assert (self.image_dimension == len(face)), \"new image not of good size\"\n",
    "\n",
    "        # Case 1: maximum number of centroids has been reached.\n",
    "        if self.centroids.shape[0] >= self.max_num_centroids + 1:\n",
    "            if self.init:\n",
    "                #  Initialization after the first time that the maximum number of centroids is reached\n",
    "                self.initialize()\n",
    "                self.init = False\n",
    "            \"\"\"\n",
    "            Find c_rep and c_add following Algorithm 1.\n",
    "            \n",
    "            - c_1, c_2 = two closest centroids (minimum distance) such that at least one of them is not in self.taboo.\n",
    "            - c_rep = centroid in {c_1, c_2} that is in self.taboo. If none of them is in self.taboo, c_rep is the one\n",
    "                      with largest multiplicity.\n",
    "            - c_add = centroid in {c_1, c_2} that is not c_rep.\n",
    "            \"\"\"\n",
    "            # get c_1, c_2\n",
    "            sorted_distances = np.argsort(self.centroids_distances.reshape(-1))\n",
    "            for ii in sorted_distances:\n",
    "                c_1 = ii // self.centroids_distances.shape[0]\n",
    "                c_2 = ii % self.centroids_distances.shape[0]\n",
    "                if c_1 not in self.taboo or c_2 not in self.taboo:\n",
    "                    break\n",
    "            # get c_rep and c_add\n",
    "            if c_1 in self.taboo:\n",
    "                c_rep = c_1\n",
    "                c_add = c_2\n",
    "            elif c_2 in self.taboo:\n",
    "                c_rep = c_2\n",
    "                c_add = c_1\n",
    "            elif self.V[c_2] <= self.V[c_1]:\n",
    "                c_rep = c_1\n",
    "                c_add = c_2\n",
    "            else:\n",
    "                c_rep = c_2\n",
    "                c_add = c_1\n",
    "            \n",
    "            \"\"\"\n",
    "            Update data structures: self.centroids and self.V\n",
    "            \"\"\"\n",
    "            self.V[c_rep] += self.V[c_add]\n",
    "            self.centroids[c_add] = face\n",
    "            self.V[c_add] = 1\n",
    "\n",
    "            \"\"\"\n",
    "            Update the matrix containing the distances.\n",
    "            \"\"\"\n",
    "            dist_row = distance.cdist(np.array([self.centroids[c_add]]), self.centroids)[0]\n",
    "            dist_row[c_add] = +np.inf\n",
    "            self.centroids_distances[c_add, :] = dist_row\n",
    "            self.centroids_distances[:, c_add] = dist_row\n",
    "            self.last_face = c_add\n",
    "\n",
    "        # Case 2: create new centroid with face\n",
    "        # Remark: the multiplicities vector self.V is initialized in self.initialize()\n",
    "        else:\n",
    "            current_len = len(self.centroids)\n",
    "            self.Y = np.append(self.Y, 0)\n",
    "            self.centroids = np.vstack([self.centroids, face])\n",
    "\n",
    "    def online_ssl_compute_solution(self):\n",
    "        \"\"\"\n",
    "        Returns a prediction corresponding to self.last_face.\n",
    "        \"\"\"\n",
    "\n",
    "        # Multiplicity matrix\n",
    "        if self.init:\n",
    "            V = np.diag(np.ones(self.centroids.shape[0]))\n",
    "            self.last_face = self.centroids.shape[0] - 1\n",
    "        else:\n",
    "            V = np.diag(self.V)\n",
    "            \n",
    "        # Build quantized graph and its regularized Laplacian        \n",
    "        W = build_similarity_graph(self.centroids, eps=params_online_ssl['eps'], \n",
    "                                   var=params_online_ssl['var'], k=params_online_ssl['k'])\n",
    "        W = V.dot(W.dot(V))\n",
    "        L = build_laplacian(W, params_online_ssl['laplacian_normalization'])\n",
    "        Q = L + params_online_ssl['laplacian_regularization'] * np.eye(L.shape[0])  # regularized Laplacian\n",
    "\n",
    "        # Compute the hard HFS solution f. \n",
    "        labels, f = compute_hfs(Q, self.Y, soft=False, **params_online_ssl)\n",
    "\n",
    "        # Return the score for each possible label\n",
    "        num_classes = len(np.unique(self.Y))-1 \n",
    "        label_scores = []\n",
    "        best_score = 0\n",
    "        for ii in range(num_classes):\n",
    "            label = self.label_names[ii]\n",
    "            score = f[self.last_face, ii]\n",
    "            label_scores.append((label, score))\n",
    "            if score>best_score:\n",
    "                best_score = score\n",
    "        # handle unknown faces\n",
    "        if best_score<0.5:\n",
    "            label_scores.append(('Unknown', 2*(0.5-best_score)))\n",
    "                \n",
    "        return label_scores\n",
    "      \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Profile found with 10 images.\n",
      "Profile found with 14 images.\n"
     ]
    }
   ],
   "source": [
    "#create_user_profile('pierre')         # choose your names here :)\n",
    "#create_user_profile('ines')\n",
    "online_face_recognition(['pierre', 'ines'], IncrementalKCenters, n_pictures=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Profile found with 11 images.\n",
      "Profile found with 13 images.\n",
      "Profile found with 11 images.\n",
      "Profile found with 13 images.\n"
     ]
    }
   ],
   "source": [
    "video_fname = 'data/yann_lex_cut.mp4'\n",
    "create_user_profile('yann', video_filename=video_fname)        \n",
    "create_user_profile('lex', video_filename=video_fname)\n",
    "online_face_recognition(['yann', 'lex'], IncrementalKCenters, n_pictures=15, video_filename=video_fname)"
   ]
  },
  {
   "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": 4
}