I had one big question after taking on this review: How relevant is this book with the advent of large language models (LLMs)? In the past two years, the launches of OpenAI’s GPT and Google’s Gemma, amongst others, have severely disrupted the study of natural language processing (NLP). Some of these models have even managed to pass bar exams [1].

This book can be separated into two parts. The first presents a comprehensive recap of many major NLP concepts, including N-gram language models, part-of-speech tagging, syntactic parsing, semantic analysis, and discourse analysis. There is also a chapter on recurrent neural networks (RNN) and transformers. The second part brings the reader through many of these concepts using Python and well-known libraries such as spaCy and NLTK.

The first part presents the content clearly, providing an in-depth look at each topic. The writing is easy to follow and the diagrams and illustrations should help most readers grasp the concepts being discussed.

The second part walks readers through how to use common Python libraries to process text, with code samples shown in a style similar to Jupyter notebooks. There are annotations along the way, and the narrative helps to explain the code to readers.

As a book seeking to present key NLP concepts developed over the past five decades, I would say that it is reasonably well done. It will definitely be helpful to readers who are interested in the field, including the theory behind the concepts, and especially those who want to learn how to work with text in a more practical way. That said, this book also feels dated given advancements in the field over the past two years. While the book does cover BERT (which was published in 2018), several important concepts, including attention, and recent LLMs have changed how the typical NLP practitioner approaches many of today’s problems.

Going back to the question I posed at the start of this review--and I might be biased given that I have spent several years working in this field--I would argue that it is still important to understand the foundations of NLP, and this book is still relevant. LLMs ultimately build on top of the N-gram language models explained in chapter 2. For anyone who is keen to learn and explore this space, this foundational knowledge is still relevant, as is a general understanding of the huge body of knowledge the community has accummulated over the past five decades.

More reviews about this item: Amazon