Coming soon

FLARERL

Low-rank Hankel structure for reinforcement learning

A research project on exploiting low-rank and Hankel structure in value functions to make deep RL agents more sample-efficient. Code, results and the paper will land here.