AAAI Publications, 2017 AAAI Fall Symposium Series

Font Size: 
EPIC Lessons for the Proposed Standard Model of the Mind
David E. Kieras

Last modified: 2017-10-09

Abstract


EPIC is a computational cognitive architecture developed for modeling human cognition and high-speed performance; on the whole, it is consistent with the proposed standard model. However, lessons learned from using EPIC to account for fundamental phenomena in the empirical psychological literature suggest that the standard model includes unnecessary constraints, in particular, the cognitive seriality and attentional bottlenecks. These standard model constraints are unjustified. They should be relaxed so that the standard model subsumes a broader class of more-plausible cognitive architectures. Two additional lessons learned from EPIC concern plausible characteristics of working-memory and functional relations between cognition, perception, and action. The proposed standard model does not take a position on these issues, and should not until the psychological theory and data are much better developed.

Keywords


cognitive architecture, human cognition, human performance

Full Text: PDF