Book

Description

Artificial intelligence (AI) is a current trend in computer science, which extends itself its amazing capacities to other technologies such as mechatronics and robotics. Obviously, AI simulates human intelligence and deep neural network which is now the essential approach of AI simulates human nervous system. AI is going beyond the knowledge-extracting ability of machine learning and its convergence is going according to two ways such as reproduction way and self-thinking way with note that reproduction fitness and self-consciousness are the two most essential faculties owned by human. Reproduction way is being represented by generative AI (GenAI) and self-thinking way is being represented by artificial general intelligence (AGI) according to my viewpoint. I cannot explain AGI yet and can only understand basic concepts about GenAI. Therefore, the book focuses on mathematical descriptions and practical explanations about GenAI, in which statistical deep generative models and transformer are two important subjects. Although how to categorize deep generative models is popular in literature books and articles, I explain carefully these models by mathematical viewpoint so that readers will comprehend logic fundamentals under interesting models. Transformer which is now the most common approaches for association of deep learning / deep neural network and self-supervised learning opens a new sky for current AI large language models like GPT and BERT but it is not easy to implement it although it is not difficult to understand it. Therefore, I try my best to describe mathematical explanations as well as practical techniques based on mathematical fundamentals for implementing transformer in the most feasible way with respect to computer programming. Moreover, a research that proposes a unified framework of deep generative models is also collected.