Transfer Learning (TL)

Transfer Learning

Project Description

Can an intelligent agent be developed which learns to play chess at a higher level or more quickly because it had previously learned to play checkers? Human beings leverage other skills all the time, and this Defense Advanced Research Projects Agency (DARPA) program worked on intelligent agents capable of similar feats. Researchers developed methods, implemented as software agents, that could transfer skills learned from a "source" task to a range of "target" tasks. Agents developed in this project may someday function as adaptive opponents in military simulations.

Accomplishments

For this project, we collaborated with the Adaptive Systems Section at NRL to investigate new techniques using case-based reinforcement learning to achieve learned performance in simulations with large decision spaces. we also built, maintained, and deployed two frameworks for evaluating intelligent agents transfer learning studies: TIELT and LIET. We extended an open-source football simulator, Rush 2008 for use in Transfer Learning research. Finally, we performed evaluations for the Transfer Learning programs using our developed tools.

Publications

Moore, P., Molineaux, M., Gupta, K.M. (2009, June 04). Casey's Quest: Transfer Learning for Adversarial Environments [Video file]. Retrieved from http://www.youtube.com/watch?v=sITkmOefamc

Laviers, K., Sukthankar, G., Molineaux, M., & Aha, D.W. (2009). Improving offensive performance through opponent modeling. To appear in Proceedings of the Fifth Conference on Artificial Intelligence and Interactive Digital Entertainment. Stanford, CA: AAAI Press.

Li, N., Stracuzzi, D.J., Cleveland, G., Langley, P., Konik, T., Shapiro, D., Ali, K., Molineaux, M., & Aha, D.W. (2009). Constructing game agents from video of human behavior. To appear in Proceedings of the Fifth Conference on Artificial Intelligence and Interactive Digital Entertainment. Stanford, CA: AAAI Press.

Aha, D.W., Molineaux, M., & Sukthankar, G. (2009). Case-based reasoning for transfer learning. Proceedings of the Eighth International Conference on Case-Based Reasoning (pp. 29-44). Seattle, WA: Springer.

Laviers, K., Sukthankar, G., Klenk, M., Aha, D.W. & Molineaux, M. (2009). Opponent modeling and spatial similarity to retrieve and reuse superior plays. In S.J. Delany (Ed.) Case-Based Reasoning for Computer Games: Papers from the ICCBR Workshop (Technical Report 7/2009). Tacoma, WA: University of Washington Tacoma, Institute of Techology. [http://gaia.fdi.ucm.es/cbrcg09]

Laviers, K., Sukthankar, G., Molineaux, M., & Aha, D.W. (2009). Exploiting early intent recognition for competitive advantage. In C. Geib, H. Bui, G. Sukthankar, & D. Pynadath (Eds.) Plan, Activity, and Intent Recognition: Papers from the IJCAI Workshop (Technical Report WS-09-28). Pasadena, CA: AAAI Press. [http://www.planrec.org/]

Li, N., Stracuzzi, D.J., Cleveland, G., Langley, P., Konik, T., Shapiro, D., Ali, K., Molineaux, M., & Aha, D.W. (2009). Learning Hierarchical Skills for Game Agents from Video of Human Behavior. In U. Kuter & H. Muñoz-Avila (Eds.) Learning Structural Knowledge from Observations: Papers from the IJCAI Workshop (Technical Report WS-09-21). Pasadena, CA: AAAI Press. [http://www.cs.umd.edu/~ukuter/struck09]

Molineaux, M., Aha, D.W., & Sukthankar, G. (2009). Beating the defense: Using plan recognition to inform learning agents. Proceedings of the Twenty-Second International FLAIRS Conference (pp. 337-343). Sanibel Island, FL: AAAI Press.

Molineaux, M., Aha, D.W., & Moore, P. (2008). Learning continuous action models in a real-time strategy environment. Proceedings of the Twenty-First Florida Artificial Intelligence Research Conference (pp. 257-262). Coconut Grove, FL: AAAI Press.