Testing a Common Model for Human and Human-Like Intelligence
Abstract
The goal of this proposal was to test the viability of the Common Model of Cognition (CMC: Figure 1) as possible systems-level brain architecture. The CMC was initially proposed by John Laird, Christian Lebiere, and Paul Rosenbloom (who served as consultants for this project) as a synthesis of decades of progress in the field of cognitively inspired AI. As such, it provides a natural computational framework to explain the human mind but remains silent as to the nature of cognition's neural bases - the human brain. Thus, we set out to examine the degree to which the CMC could provide a reasonable account of the brains intrinsic architecture. Assuming that the CMC is a valid candidate, how can its viability as a model of the human brain architecture be assessed? Operationally, a candidate model should successfully satisfy two criteria. The first is the generality criterion: the same cognitive architecture should account for brain activity data across a wide spectrum of domains and tasks. The second is the comparative superiority criterion: an ideal architecture should provide a superior fit to experimental brain data compared to competing architectures of similar complexity and generality.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 15, 2023
- Accession Number
- AD1230072
Entities
People
- Andrea Stocco
Organizations
- University of Washington