دانلود کتاب دوره‌ای در پردازش زبان طبیعی

Between 1966 and 2022, a miracle happened. 1966 was the year of the release of ELIZA, a chatbot that imitated a psychotherapist and gave the impression of leading ☞253 a conversation with its human interlocutor. When the human said, “I hate my work,” ELIZA would say, “Tell me more…,” when the human said, “I eat tons of chocolate per day,” ELIZA would say, “I think you are blocking what you really want to say,” without ever really broaching the subject. The illusion was good, but it was definitely an illusion. In 2022, ChatGPT was released. ChatGPT knows everything, has advice on everything, and can interact with humans in a perfect dialogue. Not only does it pass the Turing test, but after a few minutes, the human has completely forgotten that she is communicating with a machine. Humanity has never been so close to the technological singularity. What happened between 1966 and 2022? When did the miracle occur? In fact, there was no miracle, just hard work by thousands of people working in a discipline called Natural Language Processing or Computational Linguistics, depending on the researcher’s academic background. This book aims to introduce the reader to it. Natural Language Processing (NLP) is at the crossroads of many disciplines: linguistics, computer science, artificial intelligence, cognitive psychology, and mathematics, to name only a few. There are many books available on the subject. Every book on NLP chooses its topics among these disciplines according to its specific goals and expectations. Some books focus more on mathematics, others more on computer science, and others on linguistics. Where does this book stand in relation to its predecessors? We live in an era of transformers and embeddings, of neural architectures that accomplish fabulous things. But however fabulous these tools are, one needs to know when and why to apply them. The first step to finding meaningful solutions to a problem is to understand the problem thoroughly. This is the aim of this book. To shed light on the beautiful complexity of language. Part of this beautiful complexity consists of the fact that there are many theories, many layers, and many ways to approach language, sometimes complementary, sometimes conflicting. Some approaches to language are purely theoretical, and others involve mathematical theories or computer tools. No matter what the approach, the goal has always been to understand how language works. Theories abound, and no theory is perfect, but every theory is an opportunity to get insight into some aspect of language, knowledge, and ourselves. One needs a good technical knowledge of neural networks and a solid insight into language to apply deep learning architectures to linguistic data successfully. This book aims to give the reader insight into language, its theories, and its tools.