论文部分内容阅读
两位作者根据人工智能研究的已有成果,在1937年8月的国际人工智能会议上提出了三个基本假说:(1)知识原型——一个能有效地完成某项复杂任务的程序,必须具备相应领域的大量知识。(2)宽度假设——智能系统(程序)还应具备两种应付意外情况的能力:首先,它可以退出专业领域,转而应用一般性知识;其次,系统能够应用类比法推理,以便借用其它领域的知识。(3) 实验验证假设——人们关于人工智能的思想和方法必须在大型的问题求解系统中得到检验。这三个假说统称为知识的阈值假说。这一假说的提出,引起了学术界的广泛注意。有人认为它从更高层次上抓住了当前人工智能发展的关键问题,将给已经陷入困境的人工智能研究带来新的生机,使其产生质的飞跃,从而推动人工智能研究进入一个新的发展时期。有的则认为它只不过是一篇空泛的思辩性文章。是非曲直让读者评说,现将其精华部分译出。
Based on the existing achievements of artificial intelligence research, the authors put forward three basic hypotheses at the International Artificial Intelligence Conference, August 1937: (1) Prototype of knowledge - a procedure that can effectively accomplish a complex task must Have a lot of knowledge in the corresponding field. (2) Width Assumption - An intelligent system (program) should also have two abilities to cope with unforeseen circumstances: first, it can withdraw from the field of expertise and apply general knowledge instead; secondly, the system can apply analogical reasoning in order to borrow other Field of knowledge. (3) experimental verification hypothesis - people's ideas and methods of artificial intelligence must be tested in a large problem solving system. These three hypotheses are collectively called the threshold hypothesis of knowledge. The hypothesis raised aroused widespread attention of academics. Some people think that it grasps the key issues of the current development of artificial intelligence from a higher level, which will bring new vitality to the already-troubled artificial intelligence research and make it make a qualitative leap so as to push artificial intelligence research into a new Development period. Others think of it as nothing more than a vaguely thought-provoking article. Right and wrong let the reader commented, now the essence part of the translation.