
Natural Language Processing with TensorFlow
The definitive NLP book to implement the most sought-after machine learning models and tasks, 2nd Edition By

What’s new in this edition The code has been completely updated for TensorFlow 2. All of the chapters have been revised to use the Keras model building API rather than low level TF operations, and new examples have been added to keep up with the latest trends. Additionally, a new chapter has been added on Transformer models, and a chapter from the last edition on image captioning has been reworked to show how this can be accomplished using a machine learning pipeline with a ViT and a text-based Transformer decoder. What’s new A brand new chapter on how Transformers work and how to use a BERT model for question answering All chapters now use the Keras model building API Code revision

for TensorFlow 2 Content update to reflect new trends What trajectory does this book take to help its readers utilize TensorFlow for NLP This book balances the explanations of models and theory with practical use cases and problems. The book starts by providing a grounding in NLP and TensorFlow 2 basics and gradually builds in complexity chapter by chapter. It teaches the reader how to use word2vec, including advanced extensions, to create word embeddings that turn sequences of words into vectors accessible to deep learning algorithms. Next, chapters on classical deep learning algorithms, like CNNs and RNNs, demonstrate important NLP tasks such as sentence classification and language generation. After this the book advances to discuss high-performance RNN models, like LSTM cells and GRUs, and neural machine translation. Finally, the book explores several transformer models, and how they can be applied to NLP tasks such as question answering and image captioning. What makes TensorFlow well-suited for NLP, and how is it used in this book in particular TensorFlow is an ecosystem that supports various stages of a machine learning project, from early prototyping to productionizing the model. It can be used to solve a plethora of NLP problems which thrive on deep learning solutions. Furthermore, TensorFlow supports other features such as downloading and using pretrained deep learning models. This means TensorFlow can be vast and overwhelming, especially for novices with limited experience. This book discusses the common stages in a machine learning project, such as data exploration, feature engineering, and model training, and how they can be achieved via TensorFlow. Moreover, it discusses how to orchestrate TensorFlow with other scientific libraries such as NumPy and pandas to implement solutions during the course of projects. Key takeaways or features Learn applications of NLP across a plethora of domainsDiscover different ways to achieve the same goal using TensorFlow and
their trade-offsBuild complex data pipelines with the tf.data API for data Learn how to leverage transformer models to solve NLP problems in modern solutions
Key Topics Covered or Table of Contents Advanced Word Vector Algorithms
Sequence-to-Sequence Learning, Neural Machine Translation
Image Captioning with TransformersTransformers