Classically, imitation learning algorithms have been developed for idealized situations, e.g., the demonstrations are often required to be collected in the exact same environment and usually include ...

As president in my sophomore year, CSB tried to do a couple of things. One was social bonding—bringing everyone together through socials and having fun. The other part was sourcing career networking ...

Recent work has shown that deep neural networks are capable ofapproximating both value functions and policies in reinforcementlearning domains featuring continuous state and actionspaces. However, to ...

PhD student Yeonju Ro received the 2024 IBM PhD Fellowship award with an endowment of $40,000.

Transfer learning is a method where an agent reuses knowledge learned in a source task to improve learning on a target task. Recent work has shown that transfer learning can be extended to the idea of ...

Though computers have surpassed humans at many tasks, especially computationally intensive ones, there are many tasks for which human expertise remains necessary and/or useful. For such tasks, it is ...

Quick Index: [ A | B | C | D | E | F | G | H | I | J | K | L | M | N | O | P | Q | R | S | T | U | VWXYZ] If you are working on or know of related work not mentioned ...

Transfer Learning for Reinforcement Learning Domains: A Survey. Matthew E. Taylor and Peter Stone. Journal of Machine Learning Research, 10(1):1633–1685, 2009.

Multiagent Traffic Management: A Reservation-Based Intersection Control Mechanism. Kurt Dresner and Peter Stone. In The Third International Joint Conference on Autonomous Agents and Multiagent Systems ...

In the past five years, Isler headed research for Samsung AI Center in New York, where he helped develop AI robots for household use. On top of teaching, Isler will continue his research at UT and ...

My research interests are in the area of machine learning for speech, language, and sound processing. I am particularly interested in multimodality and unsupervised ...

Patrick MacAlpine and Peter Stone.