Book (Practical) Information science Introduction to Natural Language Processing with BERT

Product description ※Please note that product information is not in full comprehensive meaning because of the machine translation.

Informatics BERT is a natural language processing model announced by Google at the end of 2018. BERT is a natural language processing model that takes into account the "context" and achieved the highest accuracy for 11 tasks such as translation and classification, and has now established itself as a standard model. This book is an introduction to BERT, which has played a major role in the development of natural language processing in recent years and is useful in applications. In the first half of this book, we will outline natural language processing and machine learning. Then, we will actually solve various tasks with BERT. Specifically, we will deal with sentence classification, peculiar expression extraction, sentence proofreading, search for similar sentences, and visualization of data. We aim to become able to use BERT by experiencing a series of flows from data set processing to fine tuning (learning to make BERT specialize in specific language tasks) to performance evaluation. We will use Transformers as a library for processing with BERT, which is often used for deep learning language models, and PyTorch Lightning as a library for efficient learning and performance evaluation. In this book, we will systematically and carefully explain how to use Transformers and PyTorch Lightning from the beginning assuming readers who have never used Transformers or PyTorch Lightning. ▼ Environment language of this book : Python Deep Learning Framework : PyTorch Library : Transformers, PyTorch Lightning calculation environment : Google Colaboratory ▼ Features of this book : You can actually solve various tasks with BERT. ・ Data sets used are unified in Japanese. ・ We will systematically explain how to use the library from the beginning. ▼ Environment language of this book : Python Deep Learning Framework : PyTorch Library : Transformers, PyTorch Lightning calculation environment : Google Colaboratory ▼ Features of this book : You can actually solve various tasks with BERT. ・ Data sets used are unified in Japanese. ・ We will systematically explain how to use the library from the beginning.