Provably Efficient Third-Person Imitation from Offline Observation
This work addresses domain adaptation in imitation learning for improving generalizability, but it is incremental as it focuses on guarantees in a restricted setting.
The paper tackled the problem of third-person imitation learning with no strong performance guarantees, providing statistical learning guarantees for offline observation and a lower bound for online performance.
Domain adaptation in imitation learning represents an essential step towards improving generalizability. However, even in the restricted setting of third-person imitation where transfer is between isomorphic Markov Decision Processes, there are no strong guarantees on the performance of transferred policies. We present problem-dependent, statistical learning guarantees for third-person imitation from observation in an offline setting, and a lower bound on performance in the online setting.