{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Neural Networks (Feed Forwarding)\n", "\n", "**NOTE: The example and sample data is being taken from the \"Machine Learning course by Andrew Ng\" in Coursera.**" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "In the previous Example 4, we implemented multi-class logistic regression to recognize handwritten digits. However, logistic regression cannot form more complex hypotheses as it is only a linear classifier. \n", "\n", "In this Example, we will implement a neural network to recognize handwritten digits using the same training set as before. The neural network will be able to represent complex models that form non-linear hypotheses. For this Example, we will be using parameters from a neural network that we have already trained. Our goal is to implement the feedforward propagation algorithm to use our weights for prediction. In next Example, we will write the backpropagation algorithm for learning the neural network parameters." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Model representation\n", "Our neural network is shown in Figure 2. It has 3 layers – an input layer, a\n", "hidden layer and an output layer. Recall that our inputs are pixel values of\n", "digit images. Since the images are of size 20×20, this gives us 400 input layer\n", "units (excluding the extra bias unit which always outputs +1). As before,\n", "the training data will be loaded into the variables X and y.\n", "You have been provided with a set of network parameters (Θ(1) , Θ(2))\n", "already trained by us. These are stored in ex3weights.mat and will be\n", "loaded by scipy.io into Theta1 and Theta2 The parameters have dimensions\n", "that are sized for a neural network with 25 units in the second layer and 10\n", "output units (corresponding to the 10 digit classes)." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "<img src='data/nn.jpg'>" ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [], "source": [ "# initial imports\n", "import numpy as np\n", "from matplotlib import pyplot as plt\n", "%matplotlib inline\n", "import seaborn as sns" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [], "source": [ "# imports from my models\n", "from models.data_preprocessing import add_bias_unit\n", "from models.logistic_regression import cost_function, predict, gradient_descent, gradient_function, sigmoid" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Loading and Visualizing Data\n", "We start the exercise by first loading and visualizing the dataset.\n", "\n", "We will be working with a dataset that contains handwritten digits." ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [], "source": [ "# loading dataset\n", "import scipy.io as sio # sio for loading matlab file .mat\n", "data = sio.loadmat('data/ex3data1.mat')\n", "X = data['X']\n", "y = data['y']\n", "y[y==10] = 0 # mapping zeroes in y to 0" ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [], "source": [ "# setting up variables we will be using for this example\n", "input_layer_size = 400 # 20x20 Input Images of Digits\n", "hidden_layer_size = 25 # 25 hidden units\n", "num_labels = 10 # 10 labels, from 1 to 10" ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Loading and Visualizing Data ...\n", "Randomly selecting 100 data points to display\n" ] } ], "source": [ "print('Loading and Visualizing Data ...')\n", "\n", "m = X.shape[0]\n", "\n", "print(\"Randomly selecting 100 data points to display\")\n", "rand_indices = np.random.choice(range(0,m), 100)\n", "rand_samples = X[rand_indices, :]" ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "<Figure size 720x720 with 100 Axes>" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# displaying the 100 random samples using matplotlib\n", "sns.set_style('white')\n", "fig, axis = plt.subplots(10,10,sharex=True, sharey=True, figsize=(10,10))\n", "fig.subplots_adjust(wspace=0.1, hspace=0.1)\n", "axis_flt = axis.flatten()\n", "for i in range(100):\n", " axis_flt[i].imshow(rand_samples[i, :].reshape([20,20]).T, cmap='gray')" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Loading Pameters\n", "\n", "In this part of the exercise, we load some pre-initialized neural network parameters." ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Loading Saved Neural Network Parameters ...\n" ] } ], "source": [ "print('Loading Saved Neural Network Parameters ...')\n", "\n", "# Load the weights into variables Theta1 and Theta2\n", "weights = sio.loadmat('data/ex3weights.mat')\n", "theta1 = weights['Theta1'] # theta1 = numpy array of shape 25x401\n", "theta2 = weights['Theta2'] # theta2 = numpy array of shape 10x26\n", "\n", "# swap first and last columns of Theta2, due to legacy from MATLAB indexing\n", "# since the weight file ex3weights.mat was saved based on MATLAB indexing\n", "theta2 = np.roll(theta2, 1, axis=0)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Implementing Prediction" ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Accuracy of the neural network for the training set is : 97.52%\n" ] } ], "source": [ "# adding bias unit to X\n", "X = add_bias_unit(X)\n", "\n", "# predicting layer 2\n", "z2 = X @ theta1.T\n", "a2 = sigmoid(z2)\n", "\n", "# adding bias unit to a2\n", "a2 = add_bias_unit(a2)\n", "\n", "# predicting layer 3 (final layer)\n", "z3 = a2 @ theta2.T\n", "a3 = sigmoid(z3)\n", "\n", "# the predictions of final layer are saved in a3\n", "# now finding the index of largest prediction to get the most probable class\n", "\n", "pred = np.argmax(a3, axis=1).reshape(m,1)\n", "accuracy = (pred==y).astype(np.int64).mean() * 100.0\n", "print('Accuracy of the neural network for the training set is : {:.2f}%'.format(accuracy))" ] } ], "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": "3.6.8" } }, "nbformat": 4, "nbformat_minor": 4 }