近日,智源社区举行“the origin and nature of intelligence”闭门研讨会,邀请UC伯克利马毅、Doris Tsao,伦敦大学学院Karl Friston、牛津大学Michael Woorldridge、Numenta创始人Jeff Hawkins,以及智源研究院院长、北京大学黄铁军教授参与研讨。马毅教授认为:
我确实相信智能的计算和科学研究将能够很快融合。Friston教授提出了一个很好的观点,任何聪明的东西都应该具有自我好奇的关键特征。对于学习分布/相关性,我使用术语“自我批评”,对于因果关系,Judea Pearl 称之为“自我询问”,这是自主的关键。我想象一个拥有真正智能系统的新世界将是一个没有数字计算机也不需要软件的世界。


I do believe computational and scientific studies of intelligence will be able to converge soon. Prof. Karl Friston made a great point anything intelligent should have the key characteristic of being self-curious. For learning distributions/correlations, I use the term "self-critiquing", and for causality, Judea Pearl calls it "self-interrogating", a key to being autonomous. I imagine a new world with truly intelligent systems being one that there will be no digital computers and no need of softwares.
内容中包含的图片若涉及版权问题,请及时与我们联系删除


评论
沙发等你来抢