自然计算及其智能观问题  

Natural Computation and Its Problems in the View of Intelligence

作　　者：高新民[1] GAO Xinmin

机构地区：[1]华中师范大学哲学研究所,湖北武汉430079

出　　处：《上海师范大学学报（哲学社会科学版）》2020年第5期86-97,共12页Journal of Shanghai Normal University(Philosophy & Social Sciences Edition)

基　　金：教育部人文社会科学研究规划基金项目“意向性与人工智能建模基础问题研究”(20YJA720003)。

摘　　要：人工智能的自然计算研究在模拟广泛的自然事物的功能作用,建构诸如退火算法、进化算法等过程中,不仅在提升和拓展人类能力方面做出了积极的探索和贡献,而且为我们从哲学和工程技术角度进一步思考有关智能观问题、解决人工智能中深层次的基础理论问题提供了难得的素材。面对自然计算,人工智能哲学的迫切问题是区分开规范性问题和事实性问题,解决智能观中的本质循环问题。人工智能在模拟和超越自然智能上要想取得积极的进展,不能只以一种智能样式为解剖的原型实例,陷入“沙文主义”极端,而应在避免“自由主义”错误的同时放宽智能标准,坚持智能的复杂多样性理论,不犯以偏概全的错误,着力解剖尽可能多的智能样式的构成、作用方式和运作机理,以此为基础创造出更多、更有用的人工智能新样式。In the process of simulating the function of a wide range of natural things and constructing such as annealing algorithm and evolutionary algorithm,the research of natural computing of artificial intelligence not only makes active exploration and contribution in improving and expanding human ability,but also provides us with rare materials for further thinking about the problems of intelligence concept and solving the in-depth basic theoretical problems of artificial intelligence from the perspective of philosophy and engineering technology.The urgent problem of AI philosophy is to distinguish normative problem from factual problem and to solve the essential circulation problem in the view of intelligence when it comes to natural computing.If artificial intelligence wants to make positive progress in simulating and surpassing natural intelligence,it should not only take one intelligent mode as the prototype example of anatomy and fall into the“chauvinism”extreme,but also relax the intelligent standard while avoiding the“liberalism”error,adhere to the complex diversity theory of intelligence,avoid the error of over-generalization,focus on dissecting as much intelligent modes’structure and operation mechanism as possible,and base on all these work to create more and more useful new modes of artificial intelligence.

关 键 词：自然计算 智能观 本质循环问题 智能多样性 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]