COG250Y1 Lecture Notes - Lecture 13: Symbolic Artificial Intelligence, Frame Problem, Neocortex
14 Mar 2017
School
Department
Course
Professor
Document Summary
Micro worlds did not generalize to how other worlds scale up. Why did micro worlds fail: shurldu (sh) never initiated anything, it just responded to problems and situations proposed to it. This is important b/c motivation reflects what you find salient. bad. It just avoids the issue of relevance realization: sh just avoids how procedural knowledge works. It has no knowledge of this: all of the problems posed to sh are well-defined! It doesn"t have to make decisions, realize what"s relevant, anything. By this, it means that all of s( is quantitative development. S( is locked into blocks world and can"t expand its cognitive scope whatsoever. S( has no sense of the situation and therefore can"t deal with ambiguous information. Central explanation of why it failed to scale up: Sh remembers everything that happens in its environment. As soon as you try and put it in a more complex environment, every event would result in sh facing combinatorial explosion.
