Chen and Chen present a comprehensive overview of the past decades from a computer science (CS) and artificial intelligence (AI) point of view. The book is not a scientific or technical/technology elaboration since a part of it, or more exactly several parts of it, can be considered rather as an oral history. However, starting with the third part, the chapters contain lots of technical and mathematical details that describe the bases of the presented algorithms. One of the reasons that the book is worth reading--besides its encyclopedic nature--is that some chapters of technical descriptions contain intuitive and heuristic explanations of the underlying ideas in an illustrative manner, for example, in the case of gradient descent and convolution. The book is fairly long at roughly 300 pages.

The book consists of eight parts, 35 chapters, plus an appendix. The first part, “Game-Playing Machines,” overlooks the relatively recent developments of advanced systems that apply AI (machine learning, data science, computational intelligence) methods and algorithms for typically two-player games. The achievement is enormous by these systems that are generally based on the processing of huge amounts of data (big data). The result is only not unambiguous in the debater domain, where the success of arguing is decided by the audience in a subjective way. Human support for a human being is the hidden principle that is unconsciously used. The success of the machine in poker-playing promises the successful application of AI systems in an uncertain environment for decision-making.

The second part, “The Artificial Intelligence Infrastructure,” is a comprehensive overview of the history of computing and CS with personal experiences and verbal stories. It is not a systematic description of events by a historian, but rather a subjective summary.

The third part, “From Top to Bottom,” initially discusses original AI approaches, which were built on logic and rules. Furthermore, these methods provided opportunities for explanation and interpretation. Then this part analyzes the early neuron network and similar approaches that were successful in the beginning, but later led to an “AI winter.”

Part 4, “Structure and Operation,” presents the theoretical and mathematical background for a variety of artificial neuron networks (ANNs), for example, convolutional neural networks (CNNs). A full understanding of the underlying mathematics requires advanced calculus; however, the text tries to give an explanation of the principles in an illustrative way.

The fifth part, “Progression,” deals with the subtleties of some deep theoretical backgrounds, including information theory, properties of big data from the viewpoint of probability theory, parallel processing, and an outlook on quantum computing.

Part 6, “Powers of Prediction,” treatises other machine learning solutions such as the Boltzmann machine, support vector machines (SVMs), and their advantages over ANNs in specific domains, for example, in computational chemistry and the training of the machine player in video games.

Part 7, “Natural Language Processing,” showcases voice, speech recognition and processing, and the relevant theories of mathematical physics that include Fourier analysis and transformation. The treatment of the coronavirus pandemic represents an interesting application domain of the expounded techniques.

The eighth part, “The Robotworld,” addresses the issue of AI singularity and the possible superhuman capabilities of robots. Several problems are raised regarding the role of robots in war, which can be considered a reality nowadays. The successes of deep learning and relative technologies, for example, GPT3, made it possible for machines to write program code, for instance, in Python. AlphaGo Zero had no supervised training and was able to learn through “observation” and inferencing. The authors explain, however, that several results in mathematics and the theory of physics are not based on direct experimenting and observation, but on a complex process of conceptualization. Notwithstanding, there are other differences between humans and AI that are studied by philosophy, psychology, and other scientific disciplines.

The book is not technical, but rather a popular science read with occasional heavy technical details. It could be useful for people who are interested in AI, perhaps for instructors and educators who may use it to explain to students the complex technical backgrounds behind recent machine learning algorithms.