Jacob Menashe
Ph.D. Student
I am interested in model-based hierarchical reinforcement learning and monte-carlo search methods. My goal is to design learning algorithms that allow agents to balance intrinsic and extrinsic rewards in large or infinite environments with complex action dynamics. My current work seeks to do this in the context of an expansive world filled with skills, resources, and adversaries, similar to Ultima Online.
Fast and Precise Black and White Ball Detection for RoboCup Soccer 2017
Jacob Menashe, Josh Kelle, Katie Genter, Josiah Hanna, Elad Liebman, Sanmit Narvekar, Ruohan Zhang, and Peter Stone, In {R}obo{C}up-2017: Robot Soccer World Cup {XXI}, 2017 (Eds.), Nagoya, Japan, July 2017.
Monte Carlo Heirarchical Model Learning 2015
Jacob Menashe and Peter Stone, In Proceedings of the 14th International Conference on Autonomous Agents and Multiagent Systems (AAMAS), Istanbul, Turkey, May 2015.
Currently affiliated with Learning Agents