How much information should we drop to become intelligent?

被引:0
|
作者
Nave, Ophir [1 ]
Neuman, Yair [2 ]
Perlovsky, Leonid [3 ]
Howard, Newton [4 ]
机构
[1] Ben Gurion Univ Negev, Dept Math, Lev Acad Ctr JCT, IL-84105 Beer Sheva, Israel
[2] Ben Gurion Univ Negev, Dept Educ, IL-84105 Beer Sheva, Israel
[3] Harvard Univ, Sch Engn & Appl Sci, Cambridge, MA 02138 USA
[4] MIT, Synthet Intelligence Lab, Cambridge, MA 02139 USA
关键词
Cognition and physics; Set partition; Entropy; Interdisciplinary research;
D O I
10.1016/j.amc.2014.07.029
中图分类号
O29 [应用数学];
学科分类号
070104 ;
摘要
Cognitive processing by intelligent systems involves the deletion of information in favor of higher level abstractions. This process can be addressed through the physics of computation but a formal model that explains this process has not been proposed yet. In this short paper, we propose a model that through physical constraints only generates optimal solution to the collapse of n objects into n sets. A numerical simulation of the model results in a logarithmic function of information loss and condensation that perfectly fits our knowledge of cognitive processes. (C) 2014 Elsevier Inc. All rights reserved.
引用
收藏
页码:261 / 264
页数:4
相关论文
共 50 条