simple-neuralnet.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "for iteration # 0\n",
      "\n",
      "Input : \n",
      "[[0. 0. 1.]\n",
      " [0. 1. 1.]\n",
      " [1. 0. 1.]\n",
      " [1. 1. 1.]]\n",
      "Actual Output: \n",
      "[[0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]]\n",
      "Predicted Output: \n",
      "[[0.70293631]\n",
      " [0.74842117]\n",
      " [0.74724488]\n",
      " [0.77582914]]\n",
      "Loss: \n",
      "0.30580184185481446\n",
      "\n",
      "\n",
      "for iteration # 100\n",
      "\n",
      "Input : \n",
      "[[0. 0. 1.]\n",
      " [0. 1. 1.]\n",
      " [1. 0. 1.]\n",
      " [1. 1. 1.]]\n",
      "Actual Output: \n",
      "[[0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]]\n",
      "Predicted Output: \n",
      "[[0.25105129]\n",
      " [0.64100681]\n",
      " [0.58200276]\n",
      " [0.54982603]]\n",
      "Loss: \n",
      "0.16723330520988167\n",
      "\n",
      "\n",
      "for iteration # 200\n",
      "\n",
      "Input : \n",
      "[[0. 0. 1.]\n",
      " [0. 1. 1.]\n",
      " [1. 0. 1.]\n",
      " [1. 1. 1.]]\n",
      "Actual Output: \n",
      "[[0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]]\n",
      "Predicted Output: \n",
      "[[0.08459265]\n",
      " [0.83016479]\n",
      " [0.82485931]\n",
      " [0.21155173]]\n",
      "Loss: \n",
      "0.027857076827424755\n",
      "\n",
      "\n",
      "for iteration # 300\n",
      "\n",
      "Input : \n",
      "[[0. 0. 1.]\n",
      " [0. 1. 1.]\n",
      " [1. 0. 1.]\n",
      " [1. 1. 1.]]\n",
      "Actual Output: \n",
      "[[0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]]\n",
      "Predicted Output: \n",
      "[[0.04489206]\n",
      " [0.8998322 ]\n",
      " [0.89697732]\n",
      " [0.12151243]]\n",
      "Loss: \n",
      "0.009356956852487438\n",
      "\n",
      "\n",
      "for iteration # 400\n",
      "\n",
      "Input : \n",
      "[[0. 0. 1.]\n",
      " [0. 1. 1.]\n",
      " [1. 0. 1.]\n",
      " [1. 1. 1.]]\n",
      "Actual Output: \n",
      "[[0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]]\n",
      "Predicted Output: \n",
      "[[0.03178699]\n",
      " [0.92551755]\n",
      " [0.92322539]\n",
      " [0.08957169]]\n",
      "Loss: \n",
      "0.005118869262348627\n",
      "\n",
      "\n",
      "for iteration # 500\n",
      "\n",
      "Input : \n",
      "[[0. 0. 1.]\n",
      " [0. 1. 1.]\n",
      " [1. 0. 1.]\n",
      " [1. 1. 1.]]\n",
      "Actual Output: \n",
      "[[0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]]\n",
      "Predicted Output: \n",
      "[[0.02524757]\n",
      " [0.93919085]\n",
      " [0.93717949]\n",
      " [0.07276898]]\n",
      "Loss: \n",
      "0.0033942334035393395\n",
      "\n",
      "\n",
      "for iteration # 600\n",
      "\n",
      "Input : \n",
      "[[0. 0. 1.]\n",
      " [0. 1. 1.]\n",
      " [1. 0. 1.]\n",
      " [1. 1. 1.]]\n",
      "Actual Output: \n",
      "[[0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]]\n",
      "Predicted Output: \n",
      "[[0.02127266]\n",
      " [0.94785478]\n",
      " [0.94602949]\n",
      " [0.06218768]]\n",
      "Loss: \n",
      "0.0024879432797087226\n",
      "\n",
      "\n",
      "for iteration # 700\n",
      "\n",
      "Input : \n",
      "[[0. 0. 1.]\n",
      " [0. 1. 1.]\n",
      " [1. 0. 1.]\n",
      " [1. 1. 1.]]\n",
      "Actual Output: \n",
      "[[0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]]\n",
      "Predicted Output: \n",
      "[[0.01857134]\n",
      " [0.95391519]\n",
      " [0.95223007]\n",
      " [0.0548155 ]]\n",
      "Loss: \n",
      "0.0019388521649166364\n",
      "\n",
      "\n",
      "for iteration # 800\n",
      "\n",
      "Input : \n",
      "[[0. 0. 1.]\n",
      " [0. 1. 1.]\n",
      " [1. 0. 1.]\n",
      " [1. 1. 1.]]\n",
      "Actual Output: \n",
      "[[0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]]\n",
      "Predicted Output: \n",
      "[[0.01659989]\n",
      " [0.95843284]\n",
      " [0.95686038]\n",
      " [0.04933599]]\n",
      "Loss: \n",
      "0.0015746129954617902\n",
      "\n",
      "\n",
      "for iteration # 900\n",
      "\n",
      "Input : \n",
      "[[0. 0. 1.]\n",
      " [0. 1. 1.]\n",
      " [1. 0. 1.]\n",
      " [1. 1. 1.]]\n",
      "Actual Output: \n",
      "[[0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]]\n",
      "Predicted Output: \n",
      "[[0.01508832]\n",
      " [0.96195317]\n",
      " [0.96047469]\n",
      " [0.04507592]]\n",
      "Loss: \n",
      "0.0013173269522945672\n",
      "\n",
      "\n",
      "for iteration # 1000\n",
      "\n",
      "Input : \n",
      "[[0. 0. 1.]\n",
      " [0. 1. 1.]\n",
      " [1. 0. 1.]\n",
      " [1. 1. 1.]]\n",
      "Actual Output: \n",
      "[[0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]]\n",
      "Predicted Output: \n",
      "[[0.01388673]\n",
      " [0.96478755]\n",
      " [0.96338936]\n",
      " [0.04165245]]\n",
      "Loss: \n",
      "0.0011270058100791155\n",
      "\n",
      "\n",
      "for iteration # 1100\n",
      "\n",
      "Input : \n",
      "[[0. 0. 1.]\n",
      " [0. 1. 1.]\n",
      " [1. 0. 1.]\n",
      " [1. 1. 1.]]\n",
      "Actual Output: \n",
      "[[0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]]\n",
      "Predicted Output: \n",
      "[[0.01290485]\n",
      " [0.9671277 ]\n",
      " [0.96579929]\n",
      " [0.03883053]]\n",
      "Loss: \n",
      "0.0009811553608442885\n",
      "\n",
      "\n",
      "for iteration # 1200\n",
      "\n",
      "Input : \n",
      "[[0. 0. 1.]\n",
      " [0. 1. 1.]\n",
      " [1. 0. 1.]\n",
      " [1. 1. 1.]]\n",
      "Actual Output: \n",
      "[[0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]]\n",
      "Predicted Output: \n",
      "[[0.0120849 ]\n",
      " [0.96909865]\n",
      " [0.96783166]\n",
      " [0.03645719]]\n",
      "Loss: \n",
      "0.0008662168877017775\n",
      "\n",
      "\n",
      "for iteration # 1300\n",
      "\n",
      "Input : \n",
      "[[0. 0. 1.]\n",
      " [0. 1. 1.]\n",
      " [1. 0. 1.]\n",
      " [1. 1. 1.]]\n",
      "Actual Output: \n",
      "[[0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]]\n",
      "Predicted Output: \n",
      "[[0.01138808]\n",
      " [0.97078569]\n",
      " [0.96957334]\n",
      " [0.03442829]]\n",
      "Loss: \n",
      "0.0007735633188727332\n",
      "\n",
      "\n",
      "for iteration # 1400\n",
      "\n",
      "Input : \n",
      "[[0. 0. 1.]\n",
      " [0. 1. 1.]\n",
      " [1. 0. 1.]\n",
      " [1. 1. 1.]]\n",
      "Actual Output: \n",
      "[[0.]\n",
      " [1.]\n",
      " [1.]\n",
      " [0.]]\n",
      "Predicted Output: \n",
      "[[0.01078726]\n",
      " [0.97224921]\n",
      " [0.97108584]\n",
      " [0.03267024]]\n",
      "Loss: \n",
      "0.0006974611107437106\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Imports\n",
    "import numpy as np \n",
    "      \n",
    "# Each row is a training example, each column is a feature  [X1, X2, X3]\n",
    "X=np.array(([0,0,1],[0,1,1],[1,0,1],[1,1,1]), dtype=float)\n",
    "y=np.array(([0],[1],[1],[0]), dtype=float)\n",
    "\n",
    "# Define useful functions    \n",
    "\n",
    "# Activation function\n",
    "def sigmoid(t):\n",
    "    return 1/(1+np.exp(-t))\n",
    "\n",
    "# Derivative of sigmoid\n",
    "def sigmoid_derivative(p):\n",
    "    return p * (1 - p)\n",
    "\n",
    "# Class definition\n",
    "class NeuralNetwork:\n",
    "    def __init__(self, x,y):\n",
    "        self.input = x\n",
    "        self.weights1= np.random.rand(self.input.shape[1],4) # considering we have 4 nodes in the hidden layer\n",
    "        self.weights2 = np.random.rand(4,1)\n",
    "        self.y = y\n",
    "        self.output = np. zeros(y.shape)\n",
    "        \n",
    "    def feedforward(self):\n",
    "        self.layer1 = sigmoid(np.dot(self.input, self.weights1))\n",
    "        self.layer2 = sigmoid(np.dot(self.layer1, self.weights2))\n",
    "        return self.layer2\n",
    "        \n",
    "    def backprop(self):\n",
    "        d_weights2 = np.dot(self.layer1.T, 2*(self.y -self.output)*sigmoid_derivative(self.output))\n",
    "        d_weights1 = np.dot(self.input.T, np.dot(2*(self.y -self.output)*sigmoid_derivative(self.output), self.weights2.T)*sigmoid_derivative(self.layer1))\n",
    "    \n",
    "        self.weights1 += d_weights1\n",
    "        self.weights2 += d_weights2\n",
    "\n",
    "    def train(self, X, y):\n",
    "        self.output = self.feedforward()\n",
    "        self.backprop()\n",
    "        \n",
    "\n",
    "NN = NeuralNetwork(X,y)\n",
    "for i in range(1500): # trains the NN 1,000 times\n",
    "    if i % 100 ==0: \n",
    "        print (\"for iteration # \" + str(i) + \"\\n\")\n",
    "        print (\"Input : \\n\" + str(X))\n",
    "        print (\"Actual Output: \\n\" + str(y))\n",
    "        print (\"Predicted Output: \\n\" + str(NN.feedforward()))\n",
    "        print (\"Loss: \\n\" + str(np.mean(np.square(y - NN.feedforward())))) # mean sum squared loss\n",
    "        print (\"\\n\")\n",
    "  \n",
    "    NN.train(X, y)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}