دانلود کتاب پردازش زبان طبیعی با TensorFlow

This book addresses the important need for describing how Natural Language Processing (NLP) problems can be solved using TensorFlow-based NLP stacks. Deep Learning revolutionized NLP recently. Many industrial and academic NLP problems that required a large amount of work in terms of designing new features, tuning models, and finding the best modeling approach (CRF, SVM, Bayesian methods, etc.) can now be solved by NLP scientists in a significantly smaller amount of time. Also, the new deep-learning-based methods typically produce much more accurate models than traditional NLP methods. The big problem is how to make these models work in a modern production setting with operational parameters such as latency and throughput, cloud costs, and operational quality (uptime, etc.). The TensorFlow environment is designed to solve these problems when running NLP models. In this book, the author teaches the fundamentals of TensorFlow and Keras, a Python-based interface for TensorFlow. Then, the bulk of the book, from Chapter 3, Word2vec – Learning Word Embeddings, onward, is focused on NLP problems and solving them using TensorFlow. This book provides: • A knowledge of NLP methods in good detail, from their definition to various evaluation methods • Information about TensorFlow, Keras, and Hugging Face libraries, which are powerful tools to build NLP solutions • An understanding of neural architectures, which is important to build better models, by building architectures for specific tasks that the reader will encounter in their practice. The author describes the process of building embeddings and other vector representations that are the basis of most modern deep learning NLP methods. The author also describes popular Neural Network architectures, such as Recurrent Neural Networks, Convolutional Neural Networks, Long Short-Term Memory networks, and Transformer-based architectures, in detail and shows their application in solving various NLP tasks, such as sentence classification, named entity recognition, text generation, machine translation, image caption generation, and more.