Timezone: »

Offline RL Without Off-Policy Evaluation
David Brandfonbrener · Will Whitney · Rajesh Ranganath · Joan Bruna

Tue Dec 07 08:30 AM -- 10:00 AM (PST) @

Most prior approaches to offline reinforcement learning (RL) have taken an iterative actor-critic approach involving off-policy evaluation. In this paper we show that simply doing one step of constrained/regularized policy improvement using an on-policy Q estimate of the behavior policy performs surprisingly well. This one-step algorithm beats the previously reported results of iterative algorithms on a large portion of the D4RL benchmark. The one-step baseline achieves this strong performance while being notably simpler and more robust to hyperparameters than previously proposed iterative algorithms. We argue that the relatively poor performance of iterative approaches is a result of the high variance inherent in doing off-policy evaluation and magnified by the repeated optimization of policies against those estimates. In addition, we hypothesize that the strong performance of the one-step algorithm is due to a combination of favorable structure in the environment and behavior policy.

Author Information

David Brandfonbrener (New York University)
Will Whitney (DeepMind)
Rajesh Ranganath (New York University)
Joan Bruna (NYU)

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors