Bhaskara Marthi - Research Scientist
Bhaskara Marthi received his PhD in computer
science from UC Berkeley. His thesis, "Concurrent Hierarchical
Reinforcement Learning", studied the augmentation of reinforcement
learning algorithms with prior knowledge in the form of partially
specified, multithreaded programs. More recently, he has developed
hierarchical planning algorithms that intelligently combine planning at
different levels of abstraction. He has also worked on probabilistic
filtering algorithms, first-order probabilistic reasoning in worlds
with unknown objects, visual superresolution, natural language
processing, and data privacy. Dr. Marthi's current focus at Willow
Garage is on algorithms for high-level planning and state estimation in
the context of mobile manipulation.
