[1]钟义信.机制主义人工智能理论——一种通用的人工智能理论[J].智能系统学报,2018,13(1):2-18.[doi:10.11992/tis.201711032]
ZHONG Yixin.Mechanism-based artificial intelligence theory: a universal theory of artifical intelligence[J].CAAI Transactions on Intelligent Systems,2018,13(1):2-18.[doi:10.11992/tis.201711032]
点击复制
ZHONG Yixin.Mechanism-based artificial intelligence theory: a universal theory of artifical intelligence[J].CAAI Transactions on Intelligent Systems,2018,13(1):2-18.[doi:10.11992/tis.201711032]
机制主义人工智能理论——一种通用的人工智能理论
《智能系统学报》[ISSN 1673-4785/CN 23-1538/TP] 卷:
13
期数:
2018年第1期
页码:
2-18
栏目:
综述
出版日期:
2018-01-24
- Title:
- Mechanism-based artificial intelligence theory: a universal theory of artifical intelligence
- Keywords:
- methodology of information ecology; mechanism of intelligence growth; mechanism approach; general theory of artificial intelligence
- 分类号:
- TP18
- 摘要:
- 现行人工智能研究取得了许多进展，但存在“深度上浅层化、广度上碎片化和体系上封闭化”的重要缺陷。这不是改进算法或者提高硬件性能所能解决的问题，而是要在科学观方法论上寻找根源。本文依据“科学观→方法论→研究模型→研究途径→基本概念→基本原理”这个顶天立地的研究纲领，总结了信息科学的科学观，提炼了信息生态方法论；在新的科学观和方法论指导下构筑了体现智能生长全过程的研究模型，发现了智能生长的共性机制，确立了机制主义研究途径，进而澄清和匡正了信息（特别是语义信息）、感知、知识、认知、基础意识、情感、理智、综合决策等一系列基础概念，总结了实现信息－知识－智能转换的一组基本原理，创建了机制主义人工智能理论。而且证明了：长期三分而立的结构主义（人工神经网络）、功能主义（专家系统）、行为主义（感知动作系统）三大人工智能理论可在机制主义人工智能理论框架内实现和谐统一；机制主义是生成基础意识、情感、理智三位一体高等人工智能的科学途径；机制主义人工智能理论是通用型的人工智能理论。
- Abstract:
- While artificial intelligence (AI) research has made a great deal of progress, this field also faces severe drawbacks, including being shallow in depth, piece-like in width, and isolation among the pieces. These drawbacks cannot be overcome by optimizing algorithm design and improving hardware performance. The real causes of AI’s serious problems arise from its scientific view and improperly employed methodology. Following the general guidelines for scientific research architecture, i.e., “scientific view and methodology on the top, concepts and principles at the bottom, and the model and approach in between,” in this paper, we summarize the scientific view of AI and define the methodology of information ecology. In accordance with the methodology above, information research cannot stop at the information level, but must continue to the product levels-the knowledge and intelligence levels. Based on this understanding, we design a subject-object interaction model for AI research, propose a mechanism approach to AI, identify and re-build the concepts and principles underlying the entirety of information science, and establish a mechanism-based AI theory. Interestingly, we found that the three AI schools of thought that have operated in isolation from each other for decades-structuralism-based AI (the artificial neural network), functionalism-based AI (the expert system), and behaviorism-based AI (the sensor-motor system)-become harmonious components within the mechanism-based AI theory. Moreover, we also prove that the mechanism-based AI approach is also appropriate for the advanced AI work that unifies the primary consciousness, emotion, and intellect. We thus conclude that the mechanism-based AI theory provides a general theory of artificial intelligence.
- 参考文献/References:
-
[1] 钟义信. 信息科学原理[M]. 福州: 福建人民出版社, 1988.
[2] MCCULLOCH W S, PITTS W. A logical calculus of the ideas immanent in nervous activity[J]. The bulletin of mathematical biophysics, 1943, 5(4): 115-133.
[3] ROSENBLATT F. The perceptron: a probabilistic model for information storage and organization in the brain[J]. Psychological review, 1958, 65(6): 386-408.
[4] HOPFIELD J J. Neural networks and physical systems with emergent collective computational abilities[J]. Proceedings of the national academy of sciences of the United States of America, 1982, 79(8): 2554-2558.
[5] KOHONEN T. The self-organizing map[J]. Proceedings of the IEEE, 1990, 78(9): 1464-1480.
[6] KOSKO B. Adaptive bidirectional associative memories[J]. Applied optics, 1987, 26(23): 4947-4960.
[7] RUMELHART D E, MCCLELLAND J L. Parallel distributed processing[M]. Cambridge, MA, USA: MIT Press, 1986.
[8] MCCARTHY J, MINSKY M L, ROCHESTER N, et al. Proposal for the Dartmouth summer research project on artificial intelligence[R]. Technology Report. Dartmouth: Dartmouth College, 1955.
[9] SIMON H A. The sciences of the artificial[M]. Cambridge, MA, USA: MIT Press, 1969.
[10] NEWELL A. Physical symbol systems[J]. Cognitive science, 1980, 4(2): 135-183.
[11] TURING A M. Computing machinery and intelligence[J]. MIND,1950: 59, 433-460.
[12] NEWELL A, SIMON H A. GPS, a program that simulates human thought[M]//FEIGENBAUM E A, FELDMAN J. Computers and Thought. New York: McGraw-Hill, 1963: 279-293.
[13] FELGENBAUM E A. The art of artificial intelligence: themes and case studies of knowledge engineering[C]//Proceedings of the 5th International Joint Conference on Artificial Intelligence. Cambridge, USA, 1977: 1014-1029.
[14] FEIGENBAUM E A, FELDMAN J. Computers and thought[M]. New York: McGraw-Hill, 1963.
[15] NEWELL A, SIMON H A. Human problem solving[M]. Englewood Cliffs, NJ: Prentice-Hall, 1972.
[16] NILSSON N J. Principles of artificial intelligence[M]. Berlin: Springer-Verlag, 1982.
[17] MINSKY M. The society of mind[M]. New York: Simon and Schuster, 1986.
[18] BROOKS R A. Intelligence without representation[J]. Artificial intelligence, 1991, 47(1/2/3): 139-159.
[19] BROOKS R A. Elephant cannot play chess[J]. Autonomous robert,1990, 6: 3015.
[20] BROOKS R A. Engineering approach to building complete, intelligent beings[C]//Proceedings of the SPIE Intelligent Robots and Computer Vision VⅡ. Boston, MA, United States, 1989: 618-625.
[21] MINSKY M L, PAPERT S. Perceptron[M]. Cambridge: MIT Press, 1969.
[22] PAPERT S. One AI or many?[J]. Daedalus, 1988, 117(1): 1-14.
[23] RUSSELL S J, NORVIG P. Artificial intelligence: a modern approach[M]. Upper Saddle River: Pearson Education, Inc, 1995.
[24] NILSSON N J. Artificial intelligence: a new synthesis[M]. San Francisco: Morgan Kaufmann Publishers, 1998.
[25] 钟义信. 高等人工智能原理——观念·方法·模型·理论[M]. 北京: 科学出版社, 2014.
[26] 钟义信. 信息科学与技术导论[M]. 第3版. 北京: 北京邮电大学出版社, 2015.
[27] ZHONG Y X. The law of “information conversion and intelligence creation”[M]//BURGIN M HOFKIRCHNER W. Information Studies and the Quest for Transdisciplinarity. New York/London/Singapore: World Scientific, 2017: 165-190.
[28] ZHONG Y X. A theory of semantic information[J]. China communications, 2017, 14(1): 1-17.
[29] WANG P Z. A factor spaces approach to knowledge representation[J]. Fuzzy sets and systems, 1990, 36(1): 113-124.
[30] 何华灿, 王华, 刘永怀, 等. 泛逻辑学原理[M]. 北京: 科学出版社, 2001.
HE Huacan, WANG Hua, LIU Yonghuai, et al. Universal logics principle[M]. Beijing: Science Press, 2001.
[31] 何华灿, 何智涛, 王华. 论第2次数理逻辑革命[J]. 智能系统学报, 2006, 1(1): 29-37.
HE Huacan, HE Zhitao, WANG Hua. On the second revolution of mathematical logic[J]. CAAI transactions on intelligent systems, 2006, 1(1): 29-37.
- 相似文献/References:
-
[1]崔铁军,李莎莎.安全科学中的故障信息转换定律[J].智能系统学报,2020,15(2):360.[doi:10.11992/tis.201811004]
CUI Tiejun,LI Shasha.Conversion law of fault information in safety science[J].CAAI Transactions on Intelligent Systems,2020,15():360.[doi:10.11992/tis.201811004]
[2]崔铁军,李莎莎.系统运动空间与系统映射论的初步探讨[J].智能系统学报,2020,15(3):445.[doi:10.11992/tis.201902011]
CUI Tiejun,LI Shasha.Preliminary study of system movement space and system mapping theory[J].CAAI Transactions on Intelligent Systems,2020,15():445.[doi:10.11992/tis.201902011]
[3]崔铁军,李莎莎.人工智能系统故障分析原理研究[J].智能系统学报,2021,16(4):785.[doi:10.11992/tis.202003046]
CUI Tiejun,LI Shasha.Research on system fault analysis principle based on artificial intelligence system[J].CAAI Transactions on Intelligent Systems,2021,16():785.[doi:10.11992/tis.202003046]
备注/Memo
收稿日期:2017-11-28。
基金项目:国家自然科学基金项目（60873001, 60575034, 60496327, 69982001, 69171023, 68872014）.
作者简介:钟义信,男,1940年生,教授,博士生导师,主要研究方向为信息科学和人工智能的基础理论,近期主要致力于通用人工智能理论的研究。先后出版学术著作18部,发表学术论文500多篇,提出和创立了“全信息理论”、“语义信息理论”、“信息生态学理论”、“高等人工智能原理”等科学理论,多次获得国家级和部委级科技奖励。
通讯作者:钟义信.E-mail:zyx@bupt.edu.cn.
更新日期/Last Update:
2018-02-01