FARL: A Fully Automated Pipeline of RL Interface and Solver Design
Existing reinforcement learning (RL) largely focuses on optimizing learning algorithms within a pre-defined task formulation, while applying RL to new tasks still relies heavily on human expertise to design the learning pipeline: what an agent observes, how it is rewarded, which algorithm it uses, and how training is c...
Tian Jin, Liu-Yu Xiang, Chao-Wen Yang et al.
· 0 citations