Offline Imitation Learning from Multiple Baselines with Applications to Compiler Optimization

arXiv:2403.19462v1 Announce Type: new
Abstract: This work studies a Reinforcement Learning (RL) problem in which we are given a set of trajectories collected with K baseline policies. Each of these policies can be quite suboptimal in isolation, and have strong performance in complementary parts of the state space. The goal is to learn a policy which performs as well as the best combination of baselines on the entire state space. We propose a simple imitation learning based algorithm, show a …

abstract applications arxiv compiler cs.lg cs.pl imitation learning multiple offline optimization performance policies reinforcement reinforcement learning set space state studies type work