Events & News
Colloquium
| Category | Lecture |
| Date | Tuesday, March 24, 2009 |
| Time | 11:00 am |
| Location | GS 906 |
| Details | Light refreshments will be served in the 9th floor atrium at 10:45 AM |
| Speaker | Yu-Han Chang |
| Title | Research Assistant Professor |
| Affiliation | University of Southern California, ISI |
Humans and Machines in Multi-Agent Learning
AI and machine learning researchers have made great strides in
creating machines that can learn from volumes of data. But we've only
begun to explore the new possibilities that are created when rich
interaction between humans and computers is available. Humans could
instruct machines in much the same way that humans teach their own
children. The first part of the talk will introduce some work in
Wubble World that has provided glimpses of these possibilities. I'll
describe how children are able to play with their online softbot,
called a wubble, and teach the softbot about its virtual environment.
The second portion of the talk will then focus on situations where the
machines are left to fend for themselves in environments populated by
other, possibly non-cooperative, agents. These agents must learn to
adapt to, and possibly compete with, the other agents in order to
attain higher payoffs. I'll describe no-regret algorithms that can be
used in these types of scenarios to achieve good performance, even
when the agent is faced with an unknown, arbitrary opponent.
Biography
Yu-Han Chang is a Research Assistant Professor at the University of
Southern California Information Sciences Institute. His current
research interests range from reinforcement learning and game theory
to natural language understanding and interactive games. Ongoing
projects include using machine learning to improve education,
"learning by noticing", training intelligent agents via interactive
games, and developing no-regret algorithms for learning in
non-cooperative domains. Dr. Chang holds undergraduate degrees in
Mathematics and Economics, as well as a S.M. in Computer Science, from
Harvard University. He received his Ph.D. in Electrical Engineering
and Computer Science from MIT.