Mitigating Covariate Shift in Imitation Learning via Offline Data With Partial Coverage

Open in new window