Introduction to Deep Learning and Neural Networks with PythonTM: A Practical Guide

Introduction to Deep Learning and Neural Networks with Python(TM) A Practical Guide is an intensive step-by-step guide for neuroscientists to fully understand, practice, and build neural networks. Providing math and Python(TM) code examples to clarify neural network calculations, by book's end readers will fully understand how neural networks work starting from the simplest model Y=X and building from scratch. Details and explanations are provided on how a generic gradient descent algorithm works based on mathematical and Python(TM) examples, teaching you how to use the gradient descent algorithm to manually perform all calculations in both the forward and backward passes of training a neural network. Front-Matter_2021_Introduction-to-Deep-Learning-and-Neural-Networks-with-Pyt Front Matter Copyright_2021_Introduction-to-Deep-Learning-and-Neural-Networks-with-Python Copyright Dedication_2021_Introduction-to-Deep-Learning-and-Neural-Networks-with-Pytho Dedication Preface_2021_Introduction-to-Deep-Learning-and-Neural-Networks-with-Python- Preface Acknowledgment_2021_Introduction-to-Deep-Learning-and-Neural-Networks-with-P Acknowledgments Ahmed Fawzy Gad Fatima Ezzahra Jarmouni Chapter-1---Preparing-the-devel_2021_Introduction-to-Deep-Learning-and-Neura Preparing the development environment Downloading and installing PythonTM 3 Installing required libraries Preparing Ubuntu® virtual machine for Kivy Preparing Ubuntu® virtual machine for PyPy Conclusion Chapter-2---Introduction-to-artific_2021_Introduction-to-Deep-Learning-and-N Introduction to artificial neural networks (ANN) Simplest model Y = X Error calculation Introducing weight Weight as a constant Weight as a variable Optimizing the parameter Introducing bias Bias as a constant Bias as a variable Optimizing the weight and the bias From mathematical to graphical form of a neuron Neuron with multiple inputs Sum of products Activation function Conclusion Chapter-3---ANN-with-1-input_2021_Introduction-to-Deep-Learning-and-Neural-N ANN with 1 input and 1 output Network architecture Forward pass Forward pass math calculations Backward pass Chain rule Backward pass math calculations PythonTM implementation Necessary functions Preparing inputs and outputs Forward pass Backward pass Training network Conclusion Chapter-4---Working-with-any-n_2021_Introduction-to-Deep-Learning-and-Neural Working with any number of inputs ANN with 2 inputs and 1 output Math example PythonTM implementation Code changes Training ANN ANN with 10 inputs and 1 output Training ANN ANN with any number of inputs Inputs assignment Weights initialization Calculating the SOP Calculating the SOP to weights derivatives Calculating the weights gradients Updating the weights Conclusion Chapter-5---Working-with-hi_2021_Introduction-to-Deep-Learning-and-Neural-Ne Working with hidden layers ANN with 1 hidden layer with 2 neurons Forward pass Forward pass math calculations Backward pass Output layer weights Hidden layer weights Backward pass math calculations Output layer weights gradients Hidden layer weights gradients Updating weights PythonTM implementation Forward pass Backward pass Complete code Conclusion Chapter-6---Using-any-number-o_2021_Introduction-to-Deep-Learning-and-Neural Using any number of hidden neurons ANN with 1 hidden layer with 5 neurons Forward pass Backward pass Hidden layer gradients PythonTM implementation Forward pass Backward pass More iterations Any number of hidden neurons in 1 layer Weights initialization Forward pass Backward pass ANN with 8 hidden neurons Conclusion Chapter-7---Working-with-2-h_2021_Introduction-to-Deep-Learning-and-Neural-N Working with 2 hidden layers ANN with 2 hidden layers with 5 and 3 neurons Editing Chapter 6 implementation to work with an additional layer Preparing inputs, outputs, and weights Forward pass Backward pass First hidden layer gradients ANN with 2 hidden layers with 10 and 8 neurons Conclusion Chapter-8---ANN-with-3-hid_2021_Introduction-to-Deep-Learning-and-Neural-Net ANN with 3 hidden layers ANN with 3 hidden layers with 5, 3, and 2 neurons Required changes in the forward pass Required changes in the backward pass Editing Chapter 7 implementation to work with 3 hidden layers Preparing inputs, outputs, and weights Forward pass Working with any number of layers Backward pass PythonTM implementation ANN with 10 inputs and 3 hidden layers with 8, 5, and 3 neurons Conclusion Chapter-9---Working-with-any-num_2021_Introduction-to-Deep-Learning-and-Neur Working with any number of hidden layers What to do for a generic gradient descent implementation? Generic approach for gradients calculation Output layer gradients Hidden layer gradients Calculations summary PythonTM implementation backward_pass() method Output layer Hidden layers Example: Training the network Making predictions Conclusion Chapter-10---Generic_2021_Introduction-to-Deep-Learning-and-Neural-Networks- Generic ANN Preparing initial weights for any number of outputs Calculating gradients for all output neurons Network with 2 outputs Network with 3 outputs Working with multiple training samples Calculating the size of the inputs and the outputs Iterating through the training samples Calculating the network error Implementing ReLU New implementation for MLP class Example for training network with multiple samples Using bias Initializing the network bias Using bias in the forward pass Updating bias using gradient descent Complete implementation with bias Stochastic and batch gradient descent Example Conclusion Chapter-11---Running-neural-net_2021_Introduction-to-Deep-Learning-and-Neura Running neural networks in Android Building the first Kivy app Getting started with KivyMD MDTextField MDCheckbox MDDropdownMenu MDFileManager MDSpinner Training network in a thread Neural network KivyMD app neural.kv main.py Use the app Building the Android app Conclusion Index_2021_Introduction-to-Deep-Learning-and-Neural-Networks-with-Python- Index A B C D F G H I J K L M N P R S T U V W Introduction to Deep Learning and Neural Networks with PythonTM: A Practical Guide is an intensive step-by-step guide for neuroscientists to fully understand, practice, and build neural networks. Providing math and PythonTM code examples to clarify neural network calculations, by book's end readers will fully understand how neural networks work starting from the simplest model Y=X and building from scratch. Details and explanations are provided on how a generic gradient descent algorithm works based on mathematical and PythonTM examples, teaching you how to use the gradient descent algorithm to manually perform all calculations in both the forward and backward passes of training a neural network. Examines the practical side of deep learning and neural networks Provides a problem-based approach to building artificial neural networks using real data Describes PythonTM functions and features for neuroscientists Uses a careful tutorial approach to describe implementation of neural networks in PythonTM Features math and code examples (via companion website) with helpful instructions for easy implementation