Book Image

Neural Networks with Keras Cookbook

By : V Kishore Ayyadevara
Book Image

Neural Networks with Keras Cookbook

By: V Kishore Ayyadevara

Overview of this book

This book will take you from the basics of neural networks to advanced implementations of architectures using a recipe-based approach. We will learn about how neural networks work and the impact of various hyper parameters on a network's accuracy along with leveraging neural networks for structured and unstructured data. Later, we will learn how to classify and detect objects in images. We will also learn to use transfer learning for multiple applications, including a self-driving car using Convolutional Neural Networks. We will generate images while leveraging GANs and also by performing image encoding. Additionally, we will perform text analysis using word vector based techniques. Later, we will use Recurrent Neural Networks and LSTM to implement chatbot and Machine Translation systems. Finally, you will learn about transcribing images, audio, and generating captions and also use Deep Q-learning to build an agent that plays Space Invaders game. By the end of this book, you will have developed the skills to choose and customize multiple neural network architectures for various deep learning problems you might encounter.
Table of Contents (18 chapters)


In the previous chapters, we learned that the LSTM, or even the RNN, returns results from the last time step (hidden state values from the last time step are passed on to the next layer). Imagine a scenario where the output is five dimensions in size where the five dimensions are the five outputs (not softmax values for five classes). To further explain this idea, let's say we are predicting, not just the stock price on the next date, but the stock prices for the next five days. Or, we want to predict not just the next word, but a sequence of the next five words for a given combination of input sequence.

This situation calls for a different approach in building the network. In the following section, we will look into multiple scenarios of building a network to extract the outputs in different time steps.

Scenario 1: Named entity extraction

In named entity extraction...