Book Image

Hands-On Deep Learning Algorithms with Python

By : Sudharsan Ravichandiran
Book Image

Hands-On Deep Learning Algorithms with Python

By: Sudharsan Ravichandiran

Overview of this book

Deep learning is one of the most popular domains in the AI space that allows you to develop multi-layered models of varying complexities. This book introduces you to popular deep learning algorithms—from basic to advanced—and shows you how to implement them from scratch using TensorFlow. Throughout the book, you will gain insights into each algorithm, the mathematical principles involved, and how to implement it in the best possible manner. The book starts by explaining how you can build your own neural networks, followed by introducing you to TensorFlow, the powerful Python-based library for machine learning and deep learning. Moving on, you will get up to speed with gradient descent variants, such as NAG, AMSGrad, AdaDelta, Adam, and Nadam. The book will then provide you with insights into recurrent neural networks (RNNs) and LSTM and how to generate song lyrics with RNN. Next, you will master the math necessary to work with convolutional and capsule networks, widely used for image recognition tasks. You will also learn how machines understand the semantics of words and documents using CBOW, skip-gram, and PV-DM. Finally, you will explore GANs, including InfoGAN and LSGAN, and autoencoders, such as contractive autoencoders and VAE. By the end of this book, you will be equipped with all the skills you need to implement deep learning in your own projects.
Table of Contents (17 chapters)
Free Chapter
1
Section 1: Getting Started with Deep Learning
4
Section 2: Fundamental Deep Learning Algorithms
10
Section 3: Advanced Deep Learning Algorithms

Different types of RNN architectures

Now that we have learned how an RNN works, we will look at a different type of RNN architecture that's based on numbers of input and output.

One-to-one architecture

In a one-to-one architecture, a single input is mapped to a single output, and the output from the time step t is fed as an input to the next time step. We have already seen this architecture in the last section for generating songs using RNNs.

For instance, for a text generation task, we take the output generated from a current time step and feed it as the input to the next time step to generate the next word. This architecture is also widely used in stock market predictions.

The following figure shows the one-to-one RNN...