By studying humans behavior we hope to understand the limits imposed on human cognition so as to discover the appropriate constraints that can be imposed on learning algorithms. By utilizing such constraints we can simplify the problem of learning while still allowing for human like intelligence.

To achieve this goal, our lab examines two scientific problems: human vision and motor control. The vision research attempts to understand how our perception is limited by the physical limitations of the eye. This is particularly interesting because the eye is a finite resource that has to be shared between different cognitive tasks. The second direction of research that we are pursuing involves studying the characteristics of human motion. By understanding statistical properties of motion we can define a simpler space upon which learning algorithms can operate to achieve biological motion.

All of our studies are done using people and virtual reality (for vision research). Virtual reality allows us to study human behavior in natural environments, while at the same time controlling for many variables. The human behavior can then be analyzed and even replayed to generate computational models describing that behavior.