|
|
Bayesian models of perception and action Structure learning in sequential decision making Causal learning in sequential decisions made in video game environments Combining information across time in walking Models of information integration across time Learning human preferences in complex choice tasks Active information gathering Models of bistable perception in vision and audition |
|
|
SEE PUBLICATIONS |
Primary Investigator:
Postdoctoral Researchers:
Graduate Students:
Alumni:
Steve Jensen
Charles Sloane
Eric Sodomka
The goals of the lab are threefold:
1) Develop predictive models of human behavior. Specifically, the lab is interested in the integration of visual, haptic, auditory and motor information during the perception-action cycle.
2) Use pattern recognition methods that include Bayesian (Belief) Networks, Dynamic Markov Decision Networks, Pattern Theory, etc. to construct normative (optimal) solutions to perception & action problems faced by humans. The goal is to quantify the information available to a human agent and generate working solutions to computer vision problems.
3) Develop an integrated haptic (touch) and visual Virtual Reality Laboratory to explore questions about how humans use visual and haptic information to perform tasks like reaching and grasping an object.
Recent Courses
Spring 2009: CSCI 5561 Computer Vision
Fall 2008: (PSY 5993-034)
Machine Learning methods in Modeling Human learning
Equipment
Phantom Haptic Robot/VR Set-up
Used Motoman SV3X robot arm Robot arm from RobotWorx