MCCHE Precision Convergence Webinar Series with Christian Lebiere
Convergence across the Bands of Cognition
By Christian Lebiere
Carnegie Mellon University
Date: April 10, 2025
Time: 11:00 AM - 1:00 PM
Location: Online
Abstract
Human cognition involves processes and phenomena taking place at scales ranging across orders of magnitude in time and complexity that Allen Newell called the bands of cognition. He proposed the concept of cognitive architecture as a theoretical framework for integrating cognitive mechanisms across task domains and scales of activity. In this talk, I present evidence that this approach has enabled convergence within and across the bands of cognition. A variety of cognitive architectures, developed in distinct disciplines including cognitive psychology and artificial intelligence, have converged in the cognitive band onto a consensus framework called the Common Model of Cognition. Going down to the neural band, the structure of this framework and its cognitive mechanisms have been validated using neuroimaging data. Functionally, this mapping enables the development of neuro-symbolic architectures that combine the strengths of neural learning and generalization and symbolic representations and inference. Going up to the rational band, bounded rationality is enabled by reflecting the statistical regularities of the environment in the design of the cognitive mechanisms. However, systematic deviations from rationality known as cognitive biases emerge from the interaction of knowledge and processing limitations of cognitive architectures. Further up into the social band, integrating large groups of interacting cognitive agents enables the emergence of social and organizational behavior. I conclude by discussing current challenges for cognitive architectures including providing a roadmap to the convergence of human knowledge and cognitive mechanisms for the development of artificial general intelligence.
