Research Interests

My main research interest is in autonomous adaptive intelligent systems which live in the real world and improve their performance by interacting with it. I am interested in both perception and control, but most of all, in the intelligent link between the two.

The focus of my research agenda is the application of machine learning methods to problems in robotics and artificial intelligence. The central and most fascinating question I have been exploring during my graduate studies is how entities internal to the reasoning apparatus of an intelligent system (such as locations, paths, or high-level cognitive concepts) come into existence, when the immediately observable percepts consist only of sensory patterns of external influences (such as light, sound, pressure, temperature, etc.) which have no intrinsic meaning to an organism or a robot.  I believe that this question, commonly known in AI as the symbol grounding problem, holds the key to creating truly autonomous intelligent systems and explaining how mind emerges from matter.

I am a graduate of the Robotics PhD Program at Carnegie Mellon and currently work for Mitsubishi Electric Research Labs in Cambridge, Massachusetts. My dissertation work was on state-aggregation algorithms for learning probabilistic models for robot control and was supervised by Illah Nourbakhsh.