computer_vision/seminar10_1d_gan/lecture_experiments.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from scipy.stats import binom, uniform\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Биномиальное распределение "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Probability mass function values: [0.36 0.48 0.16]\n"
     ]
    },
    {
     "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": [
    "rv = binom(2, 0.4)\n",
    "x = np.arange(3)\n",
    "binom_pmf = rv.pmf(x)\n",
    "print (\"Probability mass function values: {}\".format(binom_pmf))\n",
    "plt.bar(x, binom_pmf, width=0.8, tick_label=x)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Равномерное распределение"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Probability mass function values: [0.33333333 0.33333333 0.33333333]\n"
     ]
    },
    {
     "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": [
    "uniform_pmf = 1 / len(x) * np.ones_like(x)\n",
    "print (\"Probability mass function values: {}\".format(uniform_pmf))\n",
    "plt.bar(x, uniform_pmf, width=0.8, tick_label=x)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# KL-дивергенция cвоими руками"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# https://towardsdatascience.com/kl-divergence-python-example-b87069e4b810\n",
    "def kl_divergence(p, q):\n",
    "    return np.sum(np.where(p != 0, p * np.log(p / q), 0))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "$D_\\text{KL}(P \\parallel Q)$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.0852996\n"
     ]
    }
   ],
   "source": [
    "print (\"%.7f\" % kl_divergence(binom_pmf, uniform_pmf))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "$D_\\text{KL}(Q \\parallel P)$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.097455\n"
     ]
    }
   ],
   "source": [
    "print (\"%.6f\" % kl_divergence(uniform_pmf, binom_pmf))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# KL-дивергенция из scipy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# https://docs.scipy.org/doc/scipy/reference/generated/scipy.special.kl_div.html\n",
    "from scipy.special import kl_div"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.0852996\n"
     ]
    }
   ],
   "source": [
    "print (\"%.7f\" % np.sum(kl_div(binom_pmf, uniform_pmf)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.097455\n"
     ]
    }
   ],
   "source": [
    "print (\"%.6f\" % np.sum(kl_div(uniform_pmf, binom_pmf)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# JS-дивергенция своими руками"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "def js_divergence(p, q):\n",
    "    m = (p + q) / 2\n",
    "    return (kl_divergence(p, m) + kl_divergence(q, m)) / 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.0224599\n"
     ]
    }
   ],
   "source": [
    "print (\"%.7f\" % js_divergence(binom_pmf, uniform_pmf))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.0224599\n"
     ]
    }
   ],
   "source": [
    "print (\"%.7f\" % js_divergence(uniform_pmf, binom_pmf))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# JS-дивергенция"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "# https://scipy.github.io/devdocs/generated/scipy.spatial.distance.jensenshannon.html\n",
    "from scipy.spatial.distance import jensenshannon"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Compute the Jensen-Shannon distance (metric) between two 1-D probability arrays. This is the square root  of the Jensen-Shannon divergence. The Jensen-Shannon distance between two probability vectors `p` and `q` is defined as,  $\\sqrt{\\frac{D(p \\parallel m) + D(q \\parallel m)}{2}}$, where `m` is the pointwise mean of `p` and `q` and :math:`D` is the Kullback-Leibler divergence.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.0224599\n"
     ]
    }
   ],
   "source": [
    "print (\"%.7f\" % jensenshannon(binom_pmf, uniform_pmf) ** 2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.0224599\n"
     ]
    }
   ],
   "source": [
    "print (\"%.7f\" % jensenshannon(uniform_pmf, binom_pmf) ** 2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Эксперименты с дельта-функциями"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "# https://docs.scipy.org/doc/scipy/reference/generated/scipy.signal.unit_impulse.html\n",
    "from scipy.signal import unit_impulse"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "omega = 5\n",
    "n_observations = 10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "p: [1. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      "q: [0. 0. 0. 0. 0. 1. 0. 0. 0. 0.]\n"
     ]
    }
   ],
   "source": [
    "p = unit_impulse(n_observations)\n",
    "q = unit_impulse(n_observations, idx=omega)\n",
    "print ('p: {}'.format(p))\n",
    "print ('q: {}'.format(q))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "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.vlines(np.arange(len(p)), 0, p, colors='r')\n",
    "plt.vlines(np.arange(len(p)), 0, q, colors='g')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "$D_\\text{KL}(P \\parallel Q)$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "inf"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sum(kl_div(p, q))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "$JSD (P \\parallel Q)$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6931471805599452"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "jensenshannon(p, q) ** 2 # это ln(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.allclose(np.log(2), jensenshannon(p, q) ** 2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Будем постепенно уменьшать координату в распредлении q"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "================================\n",
      "KL div: inf,\n",
      "JS div: 0.6931471805599452\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "================================\n",
      "KL div: inf,\n",
      "JS div: 0.6931471805599452\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "================================\n",
      "KL div: inf,\n",
      "JS div: 0.6931471805599452\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAD4CAYAAAD8Zh1EAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAM7ElEQVR4nO3db2xd913H8fdnycrY3yJiJEjSOYgMFk2gTlYpVIJAi5SWKXmyoUQqsKlanqzbYBWoA1RQecSGGCCFQVRGYYyGUiZmTYYgsVZIiFZx11GWhEhWtjUmRfW2UhDTyCK+PPDdeuNc+56k1z3xz++XZOmec3++/uoofuvk2NcnVYUkaeN7Rd8DSJImw6BLUiMMuiQ1wqBLUiMMuiQ1YmtfX3jbtm01PT3d15eXpA3pySef/HJVTY16rregT09PMz8/39eXl6QNKcmXVnvOSy6S1AiDLkmNMOiS1AiDLkmNMOiS1IixQU/ysSTPJfn8Ks8nye8nWUjydJK3Tn5MSdI4Xc7QHwT2rfH87cDuwcdh4KMvfSxJ0pUaG/Sq+kfgq2ssOQD8WS17HLg+yXdPakBJUjeTuIa+HTg3tL042HeZJIeTzCeZX1paurqvtnfv8oeuKXsf3MveB/f2PYa0qU0i6Bmxb+RdM6rqaFXNVNXM1NTId65Kkq7SJIK+COwc2t4BnJ/A60qSrsAkgj4L/Nzgt11uBl6oqmcn8LqSpCsw9o9zJXkI2AtsS7II/DrwSoCq+kNgDrgDWAC+BrxrvYaVJK1ubNCr6tCY5wt4z8QmkiRdFd8pKkmNMOiS1AiDLkmNMOiS1AiDLkmNMOiS1AiDLkmNMOiS1AiDLkmNMOiS1AiDLkmNMOiS1AiDLkmNMOiS1AiDLkmNMOiS1AiDLkmNMOiS1AiDLkmNMOiS1AiDLkmNMOiS1AiDLkmNMOiS1AiDLkmNMOiS1AiDLkmNMOiS1AiDLkmNMOiS1AiDLkmN6BT0JPuSnEmykOTeEc/fkOTRJE8leTrJHZMfVZK0lrFBT7IFOALcDuwBDiXZs2LZrwEPV9WNwEHgDyY9qCRpbV3O0G8CFqrqbFVdAI4BB1asKeD1g8dvAM5PbkRJUhddgr4dODe0vTjYN+w3gDuTLAJzwHtHvVCSw0nmk8wvLS1dxbiSpNV0CXpG7KsV24eAB6tqB3AH8PEkl712VR2tqpmqmpmamrryaSVJq+oS9EVg59D2Di6/pHIX8DBAVf0z8Cpg2yQGlCR10yXoJ4DdSXYluY7lH3rOrljzDHArQJI3sxx0r6lI0stobNCr6iJwN3AcOM3yb7OcTHJ/kv2DZfcA707yL8BDwDurauVlGUnSOtraZVFVzbH8w87hffcNPT4F3DLZ0SRJV8J3ikpSIwy6JDXCoEtSIwy6JDXCoEtSIwy6JDXCoEtSIwy6JDXCoEtSIwy6JDXCoEtSIwy6JDXCoEtSIwy6JDXCoEtSIwy6JDXCoEtSIwy6JDXCoEtSIwy6JDXCoEtSIwy6JDXCoEtSIwy6JDXCoEtSIwy6JDXCoEtSIwy6JDXCoEtSIwy6JDWiU9CT7EtyJslCkntXWfMzSU4lOZnkLyY7piRpnK3jFiTZAhwBfgpYBE4kma2qU0NrdgMfBG6pqueTfNd6DSxJGq3LGfpNwEJVna2qC8Ax4MCKNe8GjlTV8wBV9dxkx5QkjdMl6NuBc0Pbi4N9w94EvCnJPyV5PMm+US+U5HCS+STzS0tLVzexJGmkLkHPiH21YnsrsBvYCxwCHkhy/WWfVHW0qmaqamZqaupKZ5UkraFL0BeBnUPbO4DzI9Z8qqq+UVVfAM6wHHhJ0sukS9BPALuT7EpyHXAQmF2x5m+AnwBIso3lSzBnJzmoJGltY4NeVReBu4HjwGng4ao6meT+JPsHy44DX0lyCngU+KWq+sp6DS1JutzYX1sEqKo5YG7FvvuGHhfwgcGHJKkHvlNUkhph0CWpEQZdkhph0CWpEQZdkhph0CWpEQZdkhph0CWpEQZdkhph0CWpEQZdkhph0CWpEQZdkhph0CWpEQZdkhph0CWpEQZdkhph0CWpEQZdkhph0CWpEQZdkhph0CWpEQZdkhph0CWpEQZdkhph0CWpEQZdkhph0CWpEQZdkhph0CWpEQZdkhrRKehJ9iU5k2Qhyb1rrHt7kkoyM7kRJUldjA16ki3AEeB2YA9wKMmeEeteB7wPeGLSQ0qSxutyhn4TsFBVZ6vqAnAMODBi3W8CHwK+PsH5JEkddQn6duDc0PbiYN+3JLkR2FlVn17rhZIcTjKfZH5paemKh5Ukra5L0DNiX33ryeQVwEeAe8a9UFUdraqZqpqZmprqPqUkaawuQV8Edg5t7wDOD22/DngL8FiSLwI3A7P+YFSSXl5dgn4C2J1kV5LrgIPA7DefrKoXqmpbVU1X1TTwOLC/qubXZWJJ0khjg15VF4G7gePAaeDhqjqZ5P4k+9d7QElSN1u7LKqqOWBuxb77Vlm796WPJUm6Ur5TVJIaYdAlqREGXZIaYdAlqREGXZIaYdAlqREGXZIaYdAlqREGXZIaYdAlqREGXZIaYdAlqREGXZIaYdAlqREGXZIaYdAlqREGXZIaYdAlqREGXZIaYdAlqREGXZIaYdAlqREGXZIaYdAlqREGXZIaYdAlqREGXZIaYdAlqREGXZIaYdAlqRGdgp5kX5IzSRaS3Dvi+Q8kOZXk6ST/kOSNkx9VkrSWsUFPsgU4AtwO7AEOJdmzYtlTwExV/SDwCPChSQ8qSVpblzP0m4CFqjpbVReAY8CB4QVV9WhVfW2w+TiwY7JjSpLG6RL07cC5oe3Fwb7V3AX87agnkhxOMp9kfmlpqfuUkqSxugQ9I/bVyIXJncAM8OFRz1fV0aqaqaqZqamp7lNKksba2mHNIrBzaHsHcH7loiS3Ab8K/HhV/e9kxpMkddXlDP0EsDvJriTXAQeB2eEFSW4E/gjYX1XPTX5MSdI4Y4NeVReBu4HjwGng4ao6meT+JPsHyz4MvBb4qySfSzK7ystJktZJl0suVNUcMLdi331Dj2+b8FySpCvkO0UlqREGXZIaYdAlqREGXZIaYdAlqREGXZIaYdAlqREGXZIaYdAlqREGXZIaYdAlqREGXZIaYdAlqREGXZIaYdAlqREGXZIaYdAlqREGXZIaYdAlqREGXZIaYdAlqREGXZIaYdAlqREGXZIaYdAlqREGXZIaYdAlqREGXZIaYdAlqREGXZIaYdAlqRGdgp5kX5IzSRaS3Dvi+W9L8peD559IMj3pQSVJaxsb9CRbgCPA7cAe4FCSPSuW3QU8X1XfB3wE+K1JDypJWluXM/SbgIWqOltVF4BjwIEVaw4Afzp4/Ahwa5JMbkxJ0jhbO6zZDpwb2l4Efni1NVV1MckLwHcCXx5elOQwcBjghhtuuLqJH3vs6j5P6+qxdz7W9wjSptflDH3UmXZdxRqq6mhVzVTVzNTUVJf5JEkddQn6IrBzaHsHcH61NUm2Am8AvjqJASVJ3XQJ+glgd5JdSa4DDgKzK9bMAj8/ePx24DNVddkZuiRp/Yy9hj64Jn43cBzYAnysqk4muR+Yr6pZ4I+BjydZYPnM/OB6Di1JulyXH4pSVXPA3Ip99w09/jrwjsmOJkm6Er5TVJIaYdAlqREGXZIaYdAlqRHp67cLkywBX7rKT9/GinehbnIej0t5PF7ksbhUC8fjjVU18p2ZvQX9pUgyX1Uzfc9xrfB4XMrj8SKPxaVaPx5ecpGkRhh0SWrERg360b4HuMZ4PC7l8XiRx+JSTR+PDXkNXZJ0uY16hi5JWsGgS1IjNlzQx92werNIsjPJo0lOJzmZ5P19z3QtSLIlyVNJPt33LH1Lcn2SR5L82+DfyY/0PVNfkvzi4Pvk80keSvKqvmdaDxsq6B1vWL1ZXATuqao3AzcD79nEx2LY+4HTfQ9xjfg94O+q6geAH2KTHpck24H3ATNV9RaW/wx4k3/ie0MFnW43rN4UqurZqvrs4PF/s/zNur3fqfqVZAfw08ADfc/StySvB36M5XsVUFUXquo/+52qV1uBbx/cUe3VXH7XtSZstKCPumH1po4YQJJp4EbgiX4n6d3vAr8M/F/fg1wDvhdYAv5kcAnqgSSv6XuoPlTVvwO/DTwDPAu8UFV/3+9U62OjBb3Tzag3kySvBf4a+IWq+q++5+lLkrcBz1XVk33Pco3YCrwV+GhV3Qj8D7Apf+aU5DtY/p/8LuB7gNckubPfqdbHRgt6lxtWbxpJXslyzD9RVZ/se56e3QLsT/JFli/F/WSSP+93pF4tAotV9c3/tT3CcuA3o9uAL1TVUlV9A/gk8KM9z7QuNlrQu9ywelNIEpavj56uqt/pe56+VdUHq2pHVU2z/O/iM1XV5FlYF1X1H8C5JN8/2HUrcKrHkfr0DHBzklcPvm9updEfEHe6p+i1YrUbVvc8Vl9uAX4W+Ncknxvs+5XB/V8lgPcCnxic/JwF3tXzPL2oqieSPAJ8luXfDnuKRv8EgG/9l6RGbLRLLpKkVRh0SWqEQZekRhh0SWqEQZekRhh0SWqEQZekRvw/850MhuMq+QUAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "================================\n",
      "KL div: inf,\n",
      "JS div: 0.6931471805599452\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "================================\n",
      "KL div: inf,\n",
      "JS div: 0.6931471805599452\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "================================\n",
      "KL div: 0.0,\n",
      "JS div: 0.0\n"
     ]
    },
    {
     "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": [
    "for omega in range(5, -1, -1):\n",
    "    print ('================================')\n",
    "    p = unit_impulse(n_observations)\n",
    "    q = unit_impulse(n_observations, idx=omega)    \n",
    "    kl = np.sum(kl_div(p, q))\n",
    "    js = jensenshannon(p, q) ** 2\n",
    "    print (\"KL div: {},\\nJS div: {}\".format(kl, js))\n",
    "    plt.vlines(np.arange(len(p)), 0, p, colors='r')\n",
    "    plt.vlines(np.arange(len(p)), 0, q, colors='g')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Earth Mover's Distance (Wasserstein Distance)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "from scipy.stats import wasserstein_distance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "================================\n",
      "KL div: inf,\n",
      "JS div: 0.6931471805599452,\n",
      "EMD: 5.0\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "================================\n",
      "KL div: inf,\n",
      "JS div: 0.6931471805599452,\n",
      "EMD: 4.0\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "================================\n",
      "KL div: inf,\n",
      "JS div: 0.6931471805599452,\n",
      "EMD: 3.0\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "================================\n",
      "KL div: inf,\n",
      "JS div: 0.6931471805599452,\n",
      "EMD: 2.0\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "================================\n",
      "KL div: inf,\n",
      "JS div: 0.6931471805599452,\n",
      "EMD: 1.0\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "================================\n",
      "KL div: 0.0,\n",
      "JS div: 0.0,\n",
      "EMD: 0.0\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAD4CAYAAAD8Zh1EAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAM3klEQVR4nO3da4xc91nH8e+vdkPpNQgvEthOHYQLtSpQqlUIRIJAguQEZL8pKJbKparqN03b0AqUAgoovKJFtCCFglVKoZSEECqwKkOQaCokRCJvmhJqm0gr9+IlRdleCIiqpBEPL3baTtaznmNnNuN99vuRLM0589/ZR0f2V8dndvakqpAkbX0vmPcAkqTZMOiS1IRBl6QmDLokNWHQJamJnfP6xrt27ap9+/bN69tL0pb0yCOPfKGqFiY9N7eg79u3j6WlpXl9e0nakpJ8dqPnvOQiSU0YdElqwqBLUhMGXZKaMOiS1MTUoCf5QJInk3xqg+eT5PeSLCd5LMlrZz+mJGmaIWfoHwQOXuD5m4H9oz9Hgfc997EkSRdratCr6h+BL11gyWHgT2vNQ8CVSb5zVgNKkoaZxTX03cC5se2V0b7zJDmaZCnJ0urq6iV9sxtuv5Ibbr/ykr5WkjqbRdAzYd/Eu2ZU1bGqWqyqxYWFiZ9clSRdolkEfQXYO7a9B3hiBq8rSboIswj6ceDnRj/tch3wVFV9fgavK0m6CFN/OVeSe4AbgF1JVoBfB14IUFV/AJwAbgGWga8Ab9isYSVJG5sa9Ko6MuX5At48s4kkSZfET4pKUhMGXZKaMOiS1IRBl6QmDLokNWHQJakJgy5JTRh0SWrCoEtSEwZdkpow6JLUhEGXpCYMuiQ1YdAlqQmDLklNGHRJasKgS1ITBl2SmjDoktSEQZekJgy6JDVh0CWpCYMuSU0YdElqwqBLUhMGXZKaMOiS1IRBl6QmDLokNWHQJakJgy5JTQwKepKDSR5PspzkjgnPX5XkwSSPJnksyS2zH1WSdCFTg55kB3A3cDNwADiS5MC6Zb8G3FdV1wC3Ar8/60ElSRc25Az9WmC5qs5W1dPAvcDhdWsKePno8SuAJ2Y3oiRpiCFB3w2cG9teGe0b9xvA65OsACeAt0x6oSRHkywlWVpdXb2EcSVJGxkS9EzYV+u2jwAfrKo9wC3Ah5Kc99pVdayqFqtqcWFh4eKnlSRtaEjQV4C9Y9t7OP+SyhuB+wCq6p+BFwG7ZjGgJGmYIUE/CexPcnWSK1h70/P4ujWfA24ESPJq1oLuNRVJeh5NDXpVPQPcBjwAnGHtp1lOJbkryaHRsncAb0ryL8A9wC9U1frLMpKkTbRzyKKqOsHam53j++4ce3wauH62o0mSLoafFJWkJgy6JDVh0CWpCYMuSU0YdElqwqBLUhMGXZKaMOiS1IRBl6QmDLokNWHQJakJgy5JTRh0SWrCoEtSEwZdkpow6JLUhEGXpCYMuiQ1YdAlqQmDLklNGHRJasKgS1ITBl2SmjDoktSEQZekJgy6JDVh0CWpCYMuSU0YdElqwqBLUhODgp7kYJLHkywnuWODNT+T5HSSU0n+fLZjSpKm2TltQZIdwN3ATwArwMkkx6vq9Nia/cA7geur6stJvmOzBpYkTTbkDP1aYLmqzlbV08C9wOF1a94E3F1VXwaoqidnO6YkaZohQd8NnBvbXhntG/cq4FVJ/inJQ0kOTnqhJEeTLCVZWl1dvbSJJUkTDQl6Juyrdds7gf3ADcAR4P1Jrjzvi6qOVdViVS0uLCxc7KySpAsYEvQVYO/Y9h7giQlr/qaqvlZVnwYeZy3wkqTnyZCgnwT2J7k6yRXArcDxdWv+GvgxgCS7WLsEc3aWg0qSLmxq0KvqGeA24AHgDHBfVZ1KcleSQ6NlDwBfTHIaeBD4par64mYNLUk639QfWwSoqhPAiXX77hx7XMDbR38kSXPgJ0UlqQmDLklNGHRJasKgS1ITBl2SmjDoktSEQZekJgy6JDVh0CWpCYMuSU0YdElqwqBLUhMGXZKaMOiS1IRBl6QmDLokNWHQJakJgy5JTRh0SWrCoEtSEwZdkpow6JLUhEGXpCYMuiQ1YdAlqQmDLklNGHRJasKgS1ITBl2SmjDoktSEQZekJgYFPcnBJI8nWU5yxwXWvS5JJVmc3YiSpCGmBj3JDuBu4GbgAHAkyYEJ614GvBV4eNZDSpKmG3KGfi2wXFVnq+pp4F7g8IR1vwm8C/jqDOeTJA00JOi7gXNj2yujfd+Q5Bpgb1V99EIvlORokqUkS6urqxc9rCRpY0OCngn76htPJi8A3gO8Y9oLVdWxqlqsqsWFhYXhU0qSphoS9BVg79j2HuCJse2XAa8BPp7kM8B1wHHfGJWk59eQoJ8E9ie5OskVwK3A8a8/WVVPVdWuqtpXVfuAh4BDVbW0KRNLkiaaGvSqega4DXgAOAPcV1WnktyV5NBmDyhJGmbnkEVVdQI4sW7fnRusveG5jyVJulh+UlSSmjDoktSEQZekJgy6JDVh0CWpCYMuSU0YdElqwqBLUhMGXZKaMOiS1IRBl6QmDLokNWHQJakJgy5JTRh0SWrCoEtSEwZdkpow6JLUhEGXpCYMuiQ1YdAlqQmDLklNGHRJasKgS1ITBl2SmjDoktSEQZekJgy6JDVh0CWpCYMuSU0MCnqSg0keT7Kc5I4Jz789yekkjyX5hySvnP2okqQLmRr0JDuAu4GbgQPAkSQH1i17FFisqu8H7gfeNetBJUkXNuQM/VpguarOVtXTwL3A4fEFVfVgVX1ltPkQsGe2Y0qSphkS9N3AubHtldG+jbwR+NtJTyQ5mmQpydLq6urwKSVJUw0Jeibsq4kLk9cDi8C7Jz1fVceqarGqFhcWFoZPKUmaaueANSvA3rHtPcAT6xcluQn4VeBHq+p/ZzOeJGmoIWfoJ4H9Sa5OcgVwK3B8fEGSa4A/BA5V1ZOzH1OSNM3UoFfVM8BtwAPAGeC+qjqV5K4kh0bL3g28FPjLJJ9McnyDl5MkbZIhl1yoqhPAiXX77hx7fNOM55IkXSQ/KSpJTRh0SWrCoEtSEwZdkpow6JLUhEGXpCYMuiQ1YdAlqQmDLklNGHRJasKgS1ITBl2SmjDoktSEQZekJgy6JDVh0CWpCYMuSU0YdElqwqBLUhMGXZKaMOiS1IRBl6QmDLokNWHQJakJgy5JTRh0SWrCoEtSEwZdkpow6JLUhEGXpCYMuiQ1MSjoSQ4meTzJcpI7Jjz/LUn+YvT8w0n2zXpQSdKFTQ16kh3A3cDNwAHgSJID65a9EfhyVX0P8B7gt2Y9qCTpwoacoV8LLFfV2ap6GrgXOLxuzWHgT0aP7wduTJLZjSlJmmbngDW7gXNj2yvAD260pqqeSfIU8O3AF8YXJTkKHAW46qqrLmngj7/3Py/p6ySpuyFn6JPOtOsS1lBVx6pqsaoWFxYWhswnSRpoSNBXgL1j23uAJzZak2Qn8ArgS7MYUJI0zJCgnwT2J7k6yRXArcDxdWuOAz8/evw64GNVdd4ZuiRp80y9hj66Jn4b8ACwA/hAVZ1KchewVFXHgT8CPpRkmbUz81s3c2hJ0vmGvClKVZ0ATqzbd+fY468CPz3b0SRJF8NPikpSEwZdkpow6JLUhEGXpCYyr58uTLIKfPYSv3wX6z6Fus15PJ7N4/FNHotn63A8XllVEz+ZObegPxdJlqpqcd5zXC48Hs/m8fgmj8WzdT8eXnKRpCYMuiQ1sVWDfmzeA1xmPB7P5vH4Jo/Fs7U+HlvyGrok6Xxb9QxdkrSOQZekJrZc0KfdsHq7SLI3yYNJziQ5leRt857pcpBkR5JHk3x03rPMW5Irk9yf5N9Gf09+aN4zzUuSXxz9O/lUknuSvGjeM22GLRX0gTes3i6eAd5RVa8GrgPevI2Pxbi3AWfmPcRl4neBv6uq7wN+gG16XJLsBt4KLFbVa1j7NeAtf8X3lgo6w25YvS1U1eer6hOjx//N2j/W3fOdar6S7AF+Enj/vGeZtyQvB36EtXsVUFVPV9V2viHvTuBbR3dUezHn33Wtha0W9Ek3rN7WEQNIsg+4Bnh4vpPM3XuBXwb+b96DXAa+G1gF/nh0Cer9SV4y76Hmoar+Hfht4HPA54Gnqurv5zvV5thqQR90M+rtJMlLgb8Cbq+q/5r3PPOS5KeAJ6vqkXnPcpnYCbwWeF9VXQP8D7At33NK8m2s/U/+auC7gJckef18p9ocWy3oQ25YvW0keSFrMf9wVX1k3vPM2fXAoSSfYe1S3I8n+bP5jjRXK8BKVX39f233sxb47egm4NNVtVpVXwM+AvzwnGfaFFst6ENuWL0tJAlr10fPVNXvzHueeauqd1bVnqrax9rfi49VVcuzsCGq6j+Ac0m+d7TrRuD0HEeap88B1yV58ejfzY00fYN40D1FLxcb3bB6zmPNy/XAzwL/muSTo32/Mrr/qwTwFuDDo5Ofs8Ab5jzPXFTVw0nuBz7B2k+HPUrTXwHgR/8lqYmtdslFkrQBgy5JTRh0SWrCoEtSEwZdkpow6JLUhEGXpCb+HwQsCaYEmXceAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for omega in range(5, -1, -1):\n",
    "    print ('================================')\n",
    "    p = unit_impulse(n_observations)\n",
    "    q = unit_impulse(n_observations, idx=omega)    \n",
    "    kl = np.sum(kl_div(p, q))\n",
    "    js = jensenshannon(p, q) ** 2\n",
    "    emd = wasserstein_distance([np.argmax(p)], [np.argmax(q)]) # emd при\n",
    "    print (\"KL div: {},\\nJS div: {},\\nEMD: {}\".format(kl, js, emd))\n",
    "    plt.vlines(np.arange(len(p)), 0, p, colors='r')\n",
    "    plt.vlines(np.arange(len(p)), 0, q, colors='g')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "pytorch_1_3_1",
   "language": "python",
   "name": "pytorch_1_3_1"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}