Knowledge Transfer in Reinforcement Learning
Event Type: Seminar
Date: March 05, 2008
Time:
10:30AM
- 11:30AM
Venue:
CC-2-2540
Abstract:
Humans are remarkably good at using knowledge acquired while solving
past problems to efficiently solve novel, related problems. How can we
build artificial agents with similar capabilities? In this talk, I focus
on "reinforcement learning" (RL)---a setting where an agent must make a
sequence of decisions to reach a goal, with intermittent feedback from
the environment about the cost of its current decision. I describe an
approach that allows agents to leverage experience gained from solving
prior RL tasks. To do this, the agent learns a hierarchical Bayesian
model from previously solved RL tasks and uses it to quickly infer the
characteristics of a novel RL task. I present empirical evidence on
navigation problems and tactical battle scenarios in a real-time
strategy game, Wargus, that show that leveraging experience from prior
tasks improves the rate of convergence to a solution in a new task.
Speaker:
Soumya Ray
Speaker Bio:
Soumya Ray obtained his baccalaureate degree from the Indian Institute
of Technology, Kharagpur, and his doctorate from the University of
Wisconsin, Madison in 2005. Since 2006, he has been a postdoctoral
researcher in the machine learning group at Oregon State University. His
research interests are in statistical machine learning, reinforcement
learning and planning, and bioinformatics.
