High-Performance Artificial Intelligence Systems Based on Massive Parallelism
Finished by March 31, 2012

Search algorithms are a representative and essential method in Artificial Intelligence. They explore large search space to extract useful information. I plan to develop new techniques based on massive parallelism to dramatically speed up search algorithms. These techniques enable search to deal with much larger search space that is intractable for current techniques. I will work on games and planning as an example of my research testbed, and develop high-performance systems for these domains.