Human and machine learning across time scales
Michael Mozer
2022 Invited Talk
in
Workshop: From Cells to Societies: Collective Learning Across Scales
in
Workshop: From Cells to Societies: Collective Learning Across Scales
Abstract
People are lifelong learners who operate in natural environments with complex multiscale temporal correlations. Psychologists have argued that mechanisms of human learning and memory are optimized to the structure of the environments in which they operate. If the goal of continual learning is to design agents that interact with natural environments, we should be designing and evaluating continual learning methods for such environments.
Chat is not available.
Successful Page Load