XORdeeplearning_md.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "0fRTmjrbQwxi"
      },
      "outputs": [],
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import math\n",
        "import random\n",
        "\n",
        "def sigmoid_activation(x):\n",
        "  return (1/(1+np.exp(-x)))\n",
        "\n",
        "def derivative_sigmoid(y):\n",
        "  return ((sigmoid_activation(y)*(1-sigmoid_activation(y))))\n",
        "\n",
        "def error(target,calculated):\n",
        "  return (calculated-target)\n",
        "\n",
        "def relu(x):\n",
        "  return np.maximum(0,x)\n",
        "\n",
        "def derivative_relu(x):\n",
        "  x[x<=0]=0\n",
        "  x[x>0]=1\n",
        "  return x"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "W1=np.array([[random.random(),random.random()],[random.random(),random.random()]])\n",
        "W2=np.array([[random.random()],[random.random()]])\n",
        "biasno=random.random()\n",
        "b1=np.array([[biasno],[biasno]])\n",
        "\n",
        "X=np.array([[0.8,0.1],[0.1,0.8],[0.1,0.1],[0.8,0.8]])\n",
        "\n",
        "target=np.array([[0.9],[0.9],[0.1],[0.1]])\n",
        "changeinW2=np.zeros((2,1))\n",
        "print(changeinW2)\n",
        "changeinW1=np.zeros((2,2))\n",
        "#W2_new=np.zeros((2,2))\n",
        "#W1_new=np.zeros((2,2))\n",
        "# error_list=[]\n",
        "# xpoints_list=[]\n",
        "counter=0\n",
        "\n",
        "for m in X:\n",
        "  error_list=[]\n",
        "  xpoints_list=[]\n",
        "  for k in range (0,200000): #20085\n",
        "    H1_in=(np.dot(W1,m))+b1\n",
        "    H1_out=sigmoid_activation(H1_in)\n",
        "    H1_derivative=derivative_sigmoid(H1_in)\n",
        "\n",
        "    H2_in=(H1_out[0][0]*W2[0][0]+H1_out[1][0]*W2[1][0])\n",
        "    H2_out=sigmoid_activation(H2_in)\n",
        "    H2_derivative=derivative_sigmoid(H2_in)\n",
        "\n",
        "    TotalError=(1/2)*((error(1,H2_out))**2)\n",
        "\n",
        "\n",
        "    error_list.append(TotalError)\n",
        "    xpoints_list.append(k)\n",
        "    for i in range(0,2):\n",
        "\n",
        "      h_out=H1_out[i][0]\n",
        "      derivative = H2_derivative\n",
        "      errors=error(target[counter][0],H2_out)\n",
        "      changeinW2[i][0]=h_out*derivative*errors\n",
        "    # changeinW2=changeinW2.transpose()\n",
        "\n",
        "\n",
        "    W2=W2-(0.5*changeinW2) #learning rate=0.5\n",
        "\n",
        "    for i in range(0,2):\n",
        "      for j in range(0,2):\n",
        "        input=m[i]\n",
        "        if (j==0):\n",
        "          sum=(W2[0][0]*H2_derivative*error(target[counter][0],H2_out))\n",
        "        else:\n",
        "          sum=(W2[1][0]*H2_derivative*error(target[counter][0],H2_out))\n",
        "        changeinW1[i][j]=input*H1_derivative[j][0]*sum\n",
        "      #changeinW1[i][j]=input*H1_derivative[j][0]*((W2[i][j]*H2_derivative[j][0]*error(target[j][0],H2_out[j][0])) + (W2[i][j]*H2_derivative[j][0]*error(target[j][0],H2_out[j][0])))\n",
        "      changeinW1=changeinW1.transpose()\n",
        "\n",
        "\n",
        "    W1=W1-(0.5*changeinW1)\n",
        "\n",
        "\n",
        "\n",
        "    changeinB1=np.zeros((2,1))\n",
        "    changeinB1[0][0]=(error(target[counter][0],H2_out)*H2_derivative*(W2[0][0]*H1_derivative[0][0]+W2[1][0]*H1_derivative[1][0]))\n",
        "    changeinB1[1][0]=(error(target[counter][0],H2_out)*H2_derivative*(W2[0][0]*H1_derivative[0][0]+W2[1][0]*H1_derivative[1][0]))\n",
        "    #print(changeinB1)\n",
        "    b1=b1-(0.5*changeinB1)\n",
        "  counter+=1\n",
        "    #print(b1) #new b1\n",
        "  # W1=W1_new\n",
        "  # W2=W2_new\n",
        "  print(W1)\n",
        "  print(W2)\n",
        "  print(\"bias\",b1)\n",
        "  print(H2_out)\n",
        "  print(error_list[-1])\n",
        "  xpoints = np.array(xpoints_list)\n",
        "  ypoints = np.array(error_list)\n",
        "\n",
        "  plt.plot(xpoints, ypoints)\n",
        "  plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "0WHj-HHLQ7T9",
        "outputId": "5266ff6a-69aa-44a6-cc98-0b28cc8a7a89"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[[0.]\n",
            " [0.]]\n",
            "[[1.02351159 0.81090172]\n",
            " [0.24640067 0.70926906]]\n",
            "[[1.77844943]\n",
            " [0.9101475 ]]\n",
            "bias [[0.59784865]\n",
            " [0.59784865]]\n",
            "0.8999999999999987\n",
            "0.005000000000000131\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[[1.02469141 0.82034028]\n",
            " [0.24640067 0.71414076]]\n",
            "[[1.81068102]\n",
            " [0.94237908]]\n",
            "bias [[0.61573649]\n",
            " [0.61573649]]\n",
            "0.8999999999999987\n",
            "0.005000000000000131\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[[0.98086555 0.77651443]\n",
            " [0.24640067 0.77335028]]\n",
            "[[-1.09133666]\n",
            " [-1.9596386 ]]\n",
            "bias [[0.76957308]\n",
            " [0.76957308]]\n",
            "0.10000000000000199\n",
            "0.4049999999999982\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[[0.95753935 0.75318823]\n",
            " [0.24640067 0.72860188]]\n",
            "[[-0.80545869]\n",
            " [-1.67376062]]\n",
            "bias [[0.68447984]\n",
            " [0.68447984]]\n",
            "0.09999999999999863\n",
            "0.4050000000000012\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "W1=np.array([[0.42243681,0.44370636],[0.91999742,0.67744589]])\n",
        "W2=np.array([[-1.16120991],[-1.56432396]])\n",
        "biasno=0.73430846\n",
        "b1=np.array([[biasno],[biasno]])\n",
        "X=np.array([[0.1],[0.8]])\n",
        "\n",
        "target=np.array([[0.9]])\n",
        "H1_in=(np.dot(W1,X))+b1\n",
        "H1_out=sigmoid_activation(H1_in)\n",
        "# H1_derivative=derivative_sigmoid(H1_in)\n",
        "H2_in=(H1_out[0][0]*W2[0][0]+H1_out[1][0]*W2[1][0])\n",
        "H2_out=sigmoid_activation(H2_in)\n",
        "# H2_derivative=derivative_sigmoid(H2_in)\n",
        "TotalError=(1/2)*((error(target[0],H2_out))**2)\n",
        "print(H2_out)\n",
        "print(TotalError)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2KX56Nl4Q8Eq",
        "outputId": "f9fea175-bcce-431b-891a-5daeb10faf2a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0.1066968120676734\n",
            "[0.31466497]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import math\n",
        "import random\n",
        "\n",
        "def sigmoid_activation(x):\n",
        "  return (1/(1+np.exp(-x)))\n",
        "\n",
        "def derivative_sigmoid(y):\n",
        "  return ((sigmoid_activation(y)*(1-sigmoid_activation(y))))\n",
        "\n",
        "def error(target,calculated):\n",
        "  return (calculated-target)\n",
        "\n",
        "def relu(x):\n",
        "  return np.maximum(0,x)\n",
        "\n",
        "def derivative_relu(x):\n",
        "  # Create a copy of the input array to avoid modifying the original\n",
        "  result = np.copy(x)\n",
        "  result[result<=0] = 0\n",
        "  result[result>0] = 1\n",
        "  return result # Return the modified copy\n",
        "\n",
        "W1=np.array([[random.random(),random.random()],[random.random(),random.random()]])\n",
        "W2=np.array([[random.random()],[random.random()]])\n",
        "biasno=random.random()\n",
        "b1=np.array([[biasno],[biasno]])\n",
        "\n",
        "X=np.array([[0.8],[0.1]])\n",
        "\n",
        "target=np.array([[0.9]])\n",
        "changeinW2=np.zeros((2,1))\n",
        "print(changeinW2)\n",
        "changeinW1=np.zeros((2,2))\n",
        "#W2_new=np.zeros((2,2))\n",
        "#W1_new=np.zeros((2,2))\n",
        "error_list=[]\n",
        "xpoints_list=[]\n",
        "for k in range (0,2): #20085\n",
        "  H1_in=(np.dot(W1,X))+b1\n",
        "  H1_out=relu(H1_in)\n",
        "  H1_derivative=derivative_relu(H1_in)\n",
        "\n",
        "  H2_in=(H1_out[0][0]*W2[0][0]+H1_out[1][0]*W2[1][0])\n",
        "  H2_out=relu(H2_in)\n",
        "  # Pass a copy of H2_in to derivative_relu to avoid modifying the original\n",
        "  H2_derivative=derivative_relu(np.copy(H2_in))\n",
        "\n",
        "  TotalError=(1/2)*((error(1,H2_out))**2)\n",
        "\n",
        "\n",
        "  error_list.append(TotalError)\n",
        "  xpoints_list.append(k)\n",
        "  for i in range(0,2):\n",
        "\n",
        "    h_out=H1_out[i][0]\n",
        "    derivative = H2_derivative\n",
        "    errors=error(target,H2_out)\n",
        "    changeinW2[i][0]=h_out*derivative*errors\n",
        "  # changeinW2=changeinW2.transpose()\n",
        "\n",
        "\n",
        "  W2=W2-(0.5*changeinW2) #learning rate=0.5\n",
        "\n",
        "  for i in range(0,2):\n",
        "    for j in range(0,2):\n",
        "      input=X[i][0]\n",
        "      if (j==0):\n",
        "        sum=(W2[0][0]*H2_derivative*error(target,H2_out))\n",
        "      else:\n",
        "        sum=(W2[1][0]*H2_derivative*error(target,H2_out))\n",
        "      changeinW1[i][j]=input*H1_derivative[j][0]*sum\n",
        "    #changeinW1[i][j]=input*H1_derivative[j][0]*((W2[i][j]*H2_derivative[j][0]*error(target[j][0],H2_out[j][0])) + (W2[i][j]*H2_derivative[j][0]*error(target[j][0],H2_out[j][0])))\n",
        "    changeinW1=changeinW1.transpose()\n",
        "\n",
        "\n",
        "  W1=W1-(0.5*changeinW1)\n",
        "\n",
        "  changeinB1=np.zeros((2,1))\n",
        "  changeinB1[0][0]=(error(target,H2_out)*H2_derivative*(W2[0][0]*H1_derivative[0][0]+W2[1][0]*H1_derivative[1][0]))\n",
        "  changeinB1[1][0]=(error(target,H2_out)*H2_derivative*(W2[0][0]*H1_derivative[0][0]+W2[1][0]*H1_derivative[1][0]))\n",
        "  #print(changeinB1)\n",
        "  b1=b1-(0.5*changeinB1)\n",
        "  #print(b1) #new b1\n",
        "# W1=W1_new\n",
        "# W2=W2_new\n",
        "print(W1)\n",
        "print(W2)\n",
        "print(H2_out)\n",
        "print(error_list[-1])\n",
        "xpoints = np.array(xpoints_list)\n",
        "ypoints = np.array(error_list)\n",
        "\n",
        "plt.plot(xpoints, ypoints)\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 740
        },
        "id": "0rqxZ0y9Q_wX",
        "outputId": "60f747f8-019a-403e-c7b2-3f26fc58ccc5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[[0.]\n",
            " [0.]]\n",
            "[[0.32473646 0.36496288]\n",
            " [0.12660825 0.86913385]]\n",
            "[[0.30572369]\n",
            " [0.94392831]]\n",
            "0.8996476840489906\n",
            "0.005035293658365603\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "<ipython-input-4-47a8b7a9d561>:60: DeprecationWarning: Conversion of an array with ndim > 0 to a scalar is deprecated, and will error in future. Ensure you extract a single element from your array before performing this operation. (Deprecated NumPy 1.25.)\n",
            "  changeinW2[i][0]=h_out*derivative*errors\n",
            "<ipython-input-4-47a8b7a9d561>:73: DeprecationWarning: Conversion of an array with ndim > 0 to a scalar is deprecated, and will error in future. Ensure you extract a single element from your array before performing this operation. (Deprecated NumPy 1.25.)\n",
            "  changeinW1[i][j]=input*H1_derivative[j][0]*sum\n",
            "<ipython-input-4-47a8b7a9d561>:81: DeprecationWarning: Conversion of an array with ndim > 0 to a scalar is deprecated, and will error in future. Ensure you extract a single element from your array before performing this operation. (Deprecated NumPy 1.25.)\n",
            "  changeinB1[0][0]=(error(target,H2_out)*H2_derivative*(W2[0][0]*H1_derivative[0][0]+W2[1][0]*H1_derivative[1][0]))\n",
            "<ipython-input-4-47a8b7a9d561>:82: DeprecationWarning: Conversion of an array with ndim > 0 to a scalar is deprecated, and will error in future. Ensure you extract a single element from your array before performing this operation. (Deprecated NumPy 1.25.)\n",
            "  changeinB1[1][0]=(error(target,H2_out)*H2_derivative*(W2[0][0]*H1_derivative[0][0]+W2[1][0]*H1_derivative[1][0]))\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "s7xZQIJLROjY"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}