Artificial Intelligence

Artificial Intelligence (AI) is the study of using artificial methods and techniques to imitate, extend and expand human intelligence and realize machine intelligence. Some people divide artificial intelligence into two categories: one is symbolic intelligence and the other is computational intelligence. Symbolic intelligence is based on knowledge and problem solving through reasoning. It is also known as traditional artificial intelligence. Computational intelligence is based on data and problem solving by establishing connections through training. Artificial neural networks, genetic algorithms, fuzzy systems, evolutionary programming, artificial life, etc. can be included in computational intelligence.

Traditional artificial intelligence mainly uses knowledge for problem solving. From a practical point of view, artificial intelligence is a knowledge engineering: taking knowledge as an object, it studies the representation of knowledge, the use of knowledge and knowledge acquisition.

Artificial Intelligence has made great progress and success since it was proposed in 1956.In 1976 Newell and Simon put forward the hypothesis of a physical symbol system, which is considered to be a necessary and sufficient condition for the representation of intelligent behavior. In this way, any information processing system can be regarded as a specific physical system, such as the human nervous system, computer construction system, etc. In the 1980s, Newell et al. devoted themselves to the study of SOAR system, which is based on the theory of knowledge chunking, and utilizes the rule-based memory to obtain the search and control knowledge and operators to realize the generalized problem solving. Minsky draws from psychological research to argue that people use a large body of knowledge acquired from previous experiences and organized in their daily cognitive activities. This knowledge is memorized in the human brain in a frame-like structure. Thus, in the 1970s he proposed a frame knowledge representation. By the 1980s, Minsky believed that there was no unified theory of human intelligence, and in 1985 he published a famous book, Society of Mind. The book pointed out that the society of mind is a complex society composed of a large number of units with certain thinking ability. Represented by McCarthy and Nilsson, they advocate the use of logic to study artificial intelligence, that is, the use of formal methods to describe the objective world. The logic school emphasizes conceptualized knowledge representation, model-theoretic semantics, and deductive reasoning in AI research. McCarthy advocates that anything can be represented by a unified logical framework, centering on non-monotonic logic in common-sense reasoning. The traditional idea of artificial intelligence research is "top-down", and its goal is to make machines imitate people, thinking that the human brain's thinking activities can be defined by some formulas and rules, and therefore hoping that by translating the human way of thinking into the language of programs and inputting it into the machine, so that the machine will someday produce the same thinking ability as human beings. This theory guided early research in artificial intelligence.

Research results in neurophysiology and brain science in recent years have shown that the perceptual part of the brain, including visual, auditory, and motor cortical areas, not only functions as an input/output channel, but also has the function of directly participating in thinking. Intelligence is not only the use of knowledge and solving problems through reasoning, intelligence is also in the perceptual channel.

In 1990, Shi Zhongshi put forward a hierarchical model of human thinking, indicating that human thinking has perceptual thinking, image thinking, abstract thinking, and constitutes a hierarchical relationship. Perceptual thinking is a simple form of thinking, which produces representations through the human eye, ear, nose, tongue and body perception organs to form primary thinking. The expression of perception in perceptual thinking is the key. Figurative thinking focuses on generalization by typicalization and thinking with figurative materials that can be processed in high parallel. Abstract thinking uses physical symbol systems as a theoretical basis for expressing abstract concepts in language. The role of attention makes its processing essentially serial.