Exact Epistemology and Artificial Intelligence
- PDF / 1,488,851 Bytes
- 34 Pages / 612 x 792 pts (letter) Page_size
- 73 Downloads / 208 Views
LLIGENT SYSTEMS
Exact Epistemology and Artificial Intelligence V. K. Finna, b, * a
Federal Research Center Computer Science and Control, Russian Academy of Sciences, Moscow, 119333 Russia b Russian State University for the Humanities, Moscow, 125993 Russia *e-mail: [email protected] Received April 29, 2020
Abstract—A system of concepts and principles of exact epistemology (EE) is formulated, which is understood to be the means of acquiring new knowledge through heuristics that use an ordered set of plausible reasoning strategies. The result of their application to sequences of expandable fact bases is empirical regularities that replenish the knowledge bases of intelligent systems (ISs), which are a constructive means of achieving the principles of exact epistemology, the main concept of which is the definition of theoretical intelligence. We consider the JSM method of automated research support, which uses the interaction of induction, analogy, and abduction as a constructive method of exact epistemology for acquiring new knowledge through intelligent systems. The problems of the formation and development of artificial intelligence in Russia are discussed. Keywords: exact epistemology, induction, analogy, abduction, JSM reasoning, JSM research, empirical regularities, empirical modalities, nomological propositions, educational program for artificial intelligence DOI: 10.3103/S0005105520030073
1. EXACT EPISTEMOLOGY: ITS IDEAS AND POSTULATES The ideas and objectives of the research field of “artificial intelligence” arose in the United States in the middle of the last century in the initiative group of its creators, led by John McCarthy (with the active participation of Marvin Minsky). The aim of these researchers was the design of computer systems, whose functioning would be similar to the manifestation of the human mind. However, the birthright in the creation of such systems, whose actions of are tantamount to rational human activity, belongs to Alan Turing, who gave a positive answer to the question “Can a machine think?” and formulated his famous test in [1]. In [2], J. McCarthy and D. Hayes expressed the fundamental intention that is necessary for the development of research: artificial intelligence: “… work on artificial intelligence, especially general intelligence, will be much more successful if the very concept of intelligence is clarified.” The authors of [2] distinguish two aspects of the development of research: epistemology and heuristics. By “epistemology” they understand the formation of a system of knowledge necessary for solving problems by means of computer systems, and by “heuristic” they mean constructive means for solving problems. M. Minsky professed a different worldview, categorically stating: “What is intelligence “in reality”? From my point of view, this is more a matter of aesthetics or pride than science and technology! For me,
“intelligence” means hardly more than a set of activities that we respect, but do not understand” [3]. It should be noted that ignoring the refinement of the
Data Loading...
