Timezone: »

Near-optimal Regret for Adversarial MDP with Delayed Bandit Feedback
Tiancheng Jin · Tal Lancewicki · Haipeng Luo · Yishay Mansour · Aviv Rosenberg

Sat Jul 23 08:30 AM -- 08:50 AM (PDT) @
The standard assumption in reinforcement learning (RL) is that agents observe feedback for their actions immediately. However, in practice feedback is often observed in delay. This paper studies online learning in episodic Markov decision process (MDP) with unknown transitions, adversarially changing costs, and unrestricted delayed bandit feedback. More precisely, the feedback for the agent in episode $k$ is revealed only in the end of episode $k + d^k$, where the delay $d^k$ can be changing over episodes and chosen by an oblivious adversary. We present the first algorithms that achieve near-optimal $\sqrt{K + D}$ regret, where $K$ is the number of episodes and $D = \sum_{k=1}^K d^k$ is the total delay, significantly improving upon the best known regret bound of $(K + D)^{2/3}$.

Author Information

Tiancheng Jin (University of Southern California)
Tal Lancewicki (Tel-Aviv University)
Haipeng Luo (University of Southern California)
Yishay Mansour (Google and Tel Aviv University)
Aviv Rosenberg (Tel Aviv University)

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

More from the Same Authors