[Good book recommendation] ChatGPT entry classic "This is ChatGPT"

I. Introduction

In the current book market, there are very few books that can introduce the principles of ChatGPT in depth and broadly. However, this book manages to do just that, explaining in detail how ChatGPT works to the reader in just the right way. The goal of this book is not only to let readers understand the working principle of ChatGPT, but also to let readers have a deeper understanding and understanding of artificial intelligence language models in a simple way.

Two, easy to understand

ChatGPT is an artificial intelligence language model that works amazingly. Its ability to automatically generate human-like text is quite remarkable. How does it do it? In fact, the basic working principle of ChatGPT is to generate a "reasonable continuation" text. By "reasonable" I mean "what one might expect to see after viewing billions of human-written content on the web."

For example, suppose we have a piece of text: "The greatest advantage of artificial intelligence is its ability to...". ChatGPT scans billions of pages of human-written text, finds all instances of that text, and then sees how often the next word occurs. ChatGPT then generates a ranked list of possible successor words, along with their corresponding "probabilities".

This book explains this question step by step through short essays:

  • Add the next word according to the probability, that is to say, the output logic of ChatGPT is calculated according to the probability of the word
  • The concept of the model is derived from the probability, and there are 175 billion models that are good enough to calculate
  • Machine learning, neural network to unsupervised learning, giving examples in the dialogue allows it to learn autonomously

The detailed illustrations are also one of the reasons why this book is easy to understand. Each article has rich illustrations as a supplement to the text.

3. Legendary Boss

The author of this book, Stephen Wolfram, is also one of the highlights of this book. As a genius who can write physics books at the age of 13 and publish papers at the age of 15, the book he wrote is very popular, and was called by OpenAI CEO Sam Altman "This is the best explanation of the principle of ChatGPT that I have ever seen ".

The easy-to-understand book written by this advanced expert shows his high professional level and also reflects his educational philosophy - making complex knowledge easy to understand.

4. Wonderful introductory sequence

The highlight of this book is that the introductory preface is more exciting than the original content of the book.

There are two reasons for this. One is that the content of the introductory preface is very mellow. From the recent development trend of artificial intelligence at home and abroad, to the development history of ChatGPT, and the life experience of the author Steven Wolfram, there are detailed descriptions. Moreover, Mr. Liu Jiang, the author of the preface, was once the vice president of Beijing Zhiyuan Artificial Intelligence Research Institute, so the gold content of this preface is even higher.

Another reason is that the machine translation of the main text of this book is relatively strong, and a large number of paragraphs and sentences have not changed the word order because they were translated into Chinese. That is to say, if the reader does not always concentrate on following the context, it is easy to misunderstand what the author wants to express. I hope that the publisher can optimize the translation in the next edition so that the text is more in line with the Chinese word order.

V. Summary

In general, this book is a book that explains the principle of ChatGPT in a balanced way in terms of technology and humanities. For friends who are interested in this, it is undoubtedly a good book worth collecting. Whether you are an expert with in-depth research on artificial intelligence or a beginner who is curious about this field, you can get a wealth of knowledge and inspiration from this book

Guess you like

Origin blog.csdn.net/diandianxiyu/article/details/132038869