Book Image

Mastering NLP from Foundations to LLMs

By : Lior Gazit, Meysam Ghaffari
Book Image

Mastering NLP from Foundations to LLMs

By: Lior Gazit, Meysam Ghaffari

Overview of this book

Do you want to master Natural Language Processing (NLP) but don’t know where to begin? This book will give you the right head start. Written by leaders in machine learning and NLP, Mastering NLP from Foundations to LLMs provides an in-depth introduction to techniques. Starting with the mathematical foundations of machine learning (ML), you’ll gradually progress to advanced NLP applications such as large language models (LLMs) and AI applications. You’ll get to grips with linear algebra, optimization, probability, and statistics, which are essential for understanding and implementing machine learning and NLP algorithms. You’ll also explore general machine learning techniques and find out how they relate to NLP. Next, you’ll learn how to preprocess text data, explore methods for cleaning and preparing text for analysis, and understand how to do text classification. You’ll get all of this and more along with complete Python code samples. By the end of the book, the advanced topics of LLMs’ theory, design, and applications will be discussed along with the future trends in NLP, which will feature expert opinions. You’ll also get to strengthen your practical skills by working on sample real-world NLP business problems and solutions.
Table of Contents (14 chapters)

Challenges in developing LLMs

Developing LLMs poses a unique set of challenges, including but not limited to handling massive amounts of data, requiring vast computational resources, and the risk of introducing or perpetuating bias. The following subsections outline the detailed explanations of these challenges.

Amounts of data

LLMs require enormous amounts of data for training. As the model size grows, so does the need for diverse, high-quality training data. However, collecting and curating such large datasets is a challenging task. It can be time - consuming and expensive. There’s also the risk of inadvertently including sensitive or inappropriate data in the training set. To have more of an idea, BERT has been trained using 3.3 billion words from Wikipedia and BookCorpus. GPT-2 has been trained on 40 GB of text data, and GPT-3 has been trained on 570 GB of text data. Table 7.2 shows the number of parameters and size of training data of a few recent LMs.

...