This talk will review recent progress towards modeling how a person, animal, or machine can autonomously learn to adapt in real time to a complex and changing world. The word “autonomous” refers to the ability of such a system to learn on its own from its environment, whereas “real time” refers to the ability of the system to learn moment-by-moment, or incrementally, from its environment. This work emphasizes the system level, because a full understanding of learning requires that it be understood within a system that controls adaptive behaviors. Multi-dimensional learned information fusion plays a central role to enable the brain to combine multiple information sources to adaptively solve complex behavioral problems, and to solve outstanding technological problems that depend on intelligent and adaptive decision making and prediction in response to high-dimensional, uncertain, and ever-changing data.
Submitted: Sep 11, 2008