Aula3_Teorica_Exercicio4.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Aula3-Teorica-Exercicio4.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyM5mau1UV3H8gq4kdT2485O",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/nascimento-luciano/Artificial-Intelligence/blob/master/Aula3_Teorica_Exercicio4.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 466
        },
        "id": "HppViFJ2v2lV",
        "outputId": "018794cf-e981-40d0-8eac-39652d8cef24"
      },
      "source": [
        "import math\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "W=0;\n",
        "T=10\n",
        "x_train = np.array([-2.0, 0.0, 4.0, 1.0], dtype=float)\n",
        "y_train = np.array([2.0, 0.0, -4.0, -1.0], dtype=float)\n",
        "loss_sq=np.zeros(T)\n",
        "for t in range(0,T):\n",
        "  for n in range(0,y_train.size):\n",
        "    score=W*x_train[n]\n",
        "    residuo=score-y_train[n]\n",
        "    grad=2*residuo*x_train[n]\n",
        "    learning_rate=0.2/(np.sqrt(t+1))\n",
        "    W=W-learning_rate*grad\n",
        "  loss_sq[t]=residuo**2\n",
        "  print('loss e W=',[loss_sq[t],W])\n",
        "plt.plot(loss_sq,'k')\n",
        "plt.xlabel('interação'),plt.ylabel('Loss')\n",
        "print('W final=',W)"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "loss e W= [10.497600000000004, 0.9440000000000004]\n",
            "loss e W= [0.8106397364420599, -0.3543036072832219]\n",
            "loss e W= [0.017601455204287453, -1.1020315419935907]\n",
            "loss e W= [0.0020154603247237823, -0.9640848972182562]\n",
            "loss e W= [0.0003619333011731217, -1.0156213306876085]\n",
            "loss e W= [7.634071399873888e-05, -0.9926894806213362]\n",
            "loss e W= [1.681137734371873e-05, -1.0034802808356031]\n",
            "loss e W= [3.643060915072758e-06, -0.9983612477605066]\n",
            "loss e W= [7.511998334322153e-07, -1.0007511554709618]\n",
            "loss e W= [1.4436280714555295e-07, -0.9996681093709364]\n",
            "W final= -0.9996681093709364\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}