Book Image

Natural Language Processing with TensorFlow - Second Edition

By : Thushan Ganegedara
2 (1)
Book Image

Natural Language Processing with TensorFlow - Second Edition

2 (1)
By: Thushan Ganegedara

Overview of this book

Learning how to solve natural language processing (NLP) problems is an important skill to master due to the explosive growth of data combined with the demand for machine learning solutions in production. Natural Language Processing with TensorFlow, Second Edition, will teach you how to solve common real-world NLP problems with a variety of deep learning model architectures. The book starts by getting readers familiar with NLP and the basics of TensorFlow. Then, it gradually teaches you different facets of TensorFlow 2.x. In the following chapters, you then learn how to generate powerful word vectors, classify text, generate new text, and generate image captions, among other exciting use-cases of real-world NLP. TensorFlow has evolved to be an ecosystem that supports a machine learning workflow through ingesting and transforming data, building models, monitoring, and productionization. We will then read text directly from files and perform the required transformations through a TensorFlow data pipeline. We will also see how to use a versatile visualization tool known as TensorBoard to visualize our models. By the end of this NLP book, you will be comfortable with using TensorFlow to build deep learning models with many different architectures, and efficiently ingest data using TensorFlow Additionally, you’ll be able to confidently use TensorFlow throughout your machine learning workflow.
Table of Contents (15 chapters)
12
Other Books You May Enjoy
13
Index

How LSTMs solve the vanishing gradient problem

As we discussed earlier, even though RNNs are theoretically sound, in practice they suffer from a serious drawback. That is, when Backpropagation Through Time (BPTT) is used, the gradient diminishes quickly, which allows us to propagate the information of only a few time steps. Consequently, we can only store the information of very few time steps, thus possessing only short-term memory. This in turn limits the usefulness of RNNs in real-world sequential tasks.

Often, useful and interesting sequential tasks (such as stock market predictions or language modeling) require the ability to learn and store long-term dependencies. Think of the following example for predicting the next word:

John is a talented student. He is an A-grade student and plays rugby and cricket. All the other students envy ______.

For us, this is a very easy task. The answer would be John. However, for an RNN, this is a difficult task. We are trying to predict...