Text
E-book Artificial Intelligence Oceanography
Artificial intelligence (AI) is the core driver for the fourth technological revolution,following the revolutions in steam technology, electricity technology, and computersand information technology. Since its emergence in the 1950s, AI has fully improvedproductivity, affected and changed the production structure and production relations.UnderstandingthehistoryofAIplaysanindispensableroleinthesubsequentresearchand the development of AI technologies. AI can be divided into three generations,according to the difference in the drive mode. The subsequent subsections introduceeach of these three generations of AI. Turing proposed the “Turing Test” in 1950 [49]. It states that if a machine can answera series of questions posed by a human tester within five minutes, and more than30% of its answers can deceive the tester into thinking that they are answered bya human, then the machine can be considered intelligent. In the same year, Turingpredicted the feasibility of intelligent machines. The “Turing test” can be seen as thegenesis of AI. Newell and Simon [41] developed the first heuristic program in theworld: Logic Theorist. It successfully proved 38 theorems in the book: “Principles ofMathematics”, by simulating human thinking activities. This program successfully emonstrated the feasibility of the predictions posed by Turing, and it is consideredthe first successful AI program. In August of the same year, the concept of “artificialintelligence” was first introduced by John McCarthy, Herbert Simon, and a groupof scientists from different fields at Dartmouth College. Thus, AI stands on thestage of history as an independent discipline. Newell and Shaw [40] invented thefirst AI programming language, the information processing language (IPL). It usedsymbols as basic elements and proposed a reference table structure instead of storingaddresses or arrays. McCarthy [36] developed a list processing language based onthe IPL, which was widely used in the AI community.The first generation of AI is known as knowledge-driven AI; these AIs allowmachines to learn by imitating the process of human reasoning and thinking. AsshowninFig.1, thecoresteps canbedividedintotwoparts, knowledgerepresentationand knowledge reasoning.Knowledge representation is required to allow machines to achieve intelligentbehavior. It represents human-understood knowledge in a certain data structure thatallows machines to understand and complete the processing. The methods of knowl-edge representation include predicate logic, production rules, semantic network rep-resentation, and knowledge graphs.
Tidak tersedia versi lain