├── Activation Functions.pdf ├── README.md └── activation_functions.ipynb /Activation Functions.pdf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/kmsanjar47/Activation-Functions-Deep-Learning/d867d39e9e7f7c2361df46438dff0b3dc73768da/Activation Functions.pdf -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | # Activation Functions 2 | 3 | ## What is an activation function? 4 | 5 | While building a Neural Network, we mostly have three layers: the input layer, the hidden layer, and the output layer. The neural network has neurons that work in correspondence with weight, bias, and their respective activation function. Activation functions make back-propagation possible, allowing for the update of weights and biases based on the error at the output. 6 | 7 | The activation function decides whether a neuron should be activated by calculating the weighted sum and adding bias to it. Its purpose is to introduce non-linearity into the output of a neuron. 8 | 9 | ## Types of Activation Function: 10 | 11 | ### Sigmoid Function: 12 | 13 | It is a function plotted as an 'S' shaped graph. 14 | **Equation:** A = 1/(1 + e^(-x)) 15 | **Graph:** 16 | 17 | #### Pros: 18 | 19 | - **Smoothness:** The sigmoid function is smooth and differentiable across its entire range, aiding optimization algorithms like gradient descent. 20 | - **Output Range:** Outputs values in the range (0, 1), suitable for binary classification problems. 21 | - **Logistic Interpretation:** Closely related to logistic regression, making it convenient for binary outcomes. 22 | 23 | #### Cons: 24 | 25 | - **Vanishing Gradient Problem:** Significant issue, especially in deep neural networks. 26 | - **Output Not Centered Around Zero:** Output is not centered around zero. 27 | - **Output Saturation:** Can saturate for extreme input values. 28 | - **Not Zero-Centered:** Not zero-centered, leading to biased weight updates. 29 | 30 | ### Tanh Function: 31 | 32 | The tanh function is a mathematically shifted version of the sigmoid function. 33 | **Equation:** tanh(x) = (e^(2x) - 1) / (e^(2x) + 1) 34 | **Graph:** 35 | 36 | #### Pros: 37 | 38 | - **Zero-Centered:** Outputs range from -1 to 1, beneficial for optimization algorithms. 39 | - **Smooth and Differentiable:** Similar to the sigmoid function. 40 | - **Output Range:** Outputs values in the range (-1, 1). 41 | - **Less Likely to Saturate:** Tends to saturate less quickly than the sigmoid function. 42 | 43 | #### Cons: 44 | 45 | - **Vanishing Gradient Problem:** Still prone to the vanishing gradient problem. 46 | - **Non-Zero-Centered in Practice:** Mean of activations may shift away from zero during training. 47 | - **Computational Cost:** Involves exponentiation, computationally more expensive. 48 | - **Not Always Preferable for All Layers:** Might not be the best choice for all layers. 49 | 50 | ### ReLU Function: 51 | 52 | Stands for Rectified Linear Unit, widely used in hidden layers. 53 | **Equation:** A(x) = max(0, x) 54 | **Graph:** 55 | 56 | #### Pros: 57 | 58 | - **Simple and Efficient:** Computationally efficient with a simple threshold operation. 59 | - **Promotes Sparsity:** Tends to sparsely activate neurons, reducing overfitting. 60 | - **Mitigates Vanishing Gradient Problem:** Does not suffer from the vanishing gradient problem to the same extent. 61 | - **Accelerates Training:** Faster training of deep neural networks. 62 | - **Zero-Centered for Positive Values:** Zero-centered for positive values. 63 | 64 | #### Cons: 65 | 66 | - **Dead Neurons:** Common issue, leading to inactive neurons. 67 | - **Not Suitable for All Data:** May not be suitable for all types of data. 68 | - **Not Smooth Everywhere:** Not differentiable at zero. 69 | - **Not Always Zero-Centered:** Lack of zero-centeredness for negative values. 70 | 71 | ## References: 72 | Geeks-for-geeks & Google 73 | Picture collected from Geeks-for-geeks 74 | -------------------------------------------------------------------------------- /activation_functions.ipynb: -------------------------------------------------------------------------------- 1 | { 2 | "cells": [ 3 | { 4 | "cell_type": "code", 5 | "execution_count": 1, 6 | "id": "32af7ce9-7d7d-4f4d-adc5-74eabac4a44c", 7 | "metadata": {}, 8 | "outputs": [], 9 | "source": [ 10 | "import numpy as np\n", 11 | "import matplotlib.pyplot as plt" 12 | ] 13 | }, 14 | { 15 | "cell_type": "code", 16 | "execution_count": 2, 17 | "id": "de2c4a83-a1bd-4cf9-89b6-e3158e7e2686", 18 | "metadata": {}, 19 | "outputs": [], 20 | "source": [ 21 | "def sigmoid(x):\n", 22 | " return 1 / (1 + np.exp(-x))\n", 23 | "\n", 24 | "def tanh(x):\n", 25 | " return np.tanh(x)\n", 26 | "\n", 27 | "def relu(x):\n", 28 | " return np.maximum(0, x)" 29 | ] 30 | }, 31 | { 32 | "cell_type": "code", 33 | "execution_count": 3, 34 | "id": "4ea7c0da-2d0f-45aa-8a4e-2329506449f4", 35 | "metadata": {}, 36 | "outputs": [], 37 | "source": [ 38 | "X = np.linspace(-3, 3, 100)" 39 | ] 40 | }, 41 | { 42 | "cell_type": "code", 43 | "execution_count": 4, 44 | "id": "16399ea1-0147-444a-8d52-454fbb6f00c5", 45 | "metadata": {}, 46 | "outputs": [], 47 | "source": [ 48 | "y_sigmoid = sigmoid(X)\n", 49 | "y_tanh = tanh(X)\n", 50 | "y_relu = relu(X)" 51 | ] 52 | }, 53 | { 54 | "cell_type": "code", 55 | "execution_count": 28, 56 | "id": "90912f48-776c-4049-8c9f-0b232c5c60f2", 57 | "metadata": {}, 58 | "outputs": [ 59 | { 60 | "data": { 61 | "image/png": "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", 62 | "text/plain": [ 63 | "
" 64 | ] 65 | }, 66 | "metadata": {}, 67 | "output_type": "display_data" 68 | } 69 | ], 70 | "source": [ 71 | "plt.plot(X, y_sigmoid)\n", 72 | "plt.grid(True)" 73 | ] 74 | }, 75 | { 76 | "cell_type": "code", 77 | "execution_count": 29, 78 | "id": "a17c1b9b-65ea-471c-9eb8-4ba587287ca1", 79 | "metadata": {}, 80 | "outputs": [ 81 | { 82 | "data": { 83 | "image/png": "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", 84 | "text/plain": [ 85 | "
" 86 | ] 87 | }, 88 | "metadata": {}, 89 | "output_type": "display_data" 90 | } 91 | ], 92 | "source": [ 93 | "plt.plot(X, y_tanh)\n", 94 | "plt.grid(True)" 95 | ] 96 | }, 97 | { 98 | "cell_type": "code", 99 | "execution_count": 30, 100 | "id": "c7252edf-69e5-48a6-869b-7b0f999a4846", 101 | "metadata": {}, 102 | "outputs": [ 103 | { 104 | "data": { 105 | "image/png": "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", 106 | "text/plain": [ 107 | "
" 108 | ] 109 | }, 110 | "metadata": {}, 111 | "output_type": "display_data" 112 | } 113 | ], 114 | "source": [ 115 | "plt.plot(X, y_relu)\n", 116 | "plt.grid(True)" 117 | ] 118 | }, 119 | { 120 | "cell_type": "code", 121 | "execution_count": null, 122 | "id": "b3a89c5e-243a-41e8-9c1b-f48fccde4f49", 123 | "metadata": {}, 124 | "outputs": [], 125 | "source": [] 126 | } 127 | ], 128 | "metadata": { 129 | "kernelspec": { 130 | "display_name": "Python 3 (ipykernel)", 131 | "language": "python", 132 | "name": "python3" 133 | }, 134 | "language_info": { 135 | "codemirror_mode": { 136 | "name": "ipython", 137 | "version": 3 138 | }, 139 | "file_extension": ".py", 140 | "mimetype": "text/x-python", 141 | "name": "python", 142 | "nbconvert_exporter": "python", 143 | "pygments_lexer": "ipython3", 144 | "version": "3.11.5" 145 | } 146 | }, 147 | "nbformat": 4, 148 | "nbformat_minor": 5 149 | } 150 | --------------------------------------------------------------------------------