Beyond Value-function Gaps: Improved Instance-dependent Regret Bounds For Episodic Reinforcement Learning

Abstract

We provide improved gap-dependent regret bounds for reinforcement learning in finite episodic Markov decision processes. Compared to prior work, our bounds depend on alternative definitions of gaps. These definitions are based on the insight that, in order to achieve a favorable regret, an algorithm does not need to learn how to behave optimally in states that are not reached by an optimal policy. We prove tighter upper regret bounds for optimistic algorithms and accompany them with new information-theoretic lower bounds for a large class of MDPs. Our results show that optimistic algorithms can not achieve the information-theoretic lower bounds even in deterministic MDPs unless there is a unique optimal policy.

Stats

• citations0
• S2 citations—
• github stars0
• HF likes0
• heat score0.00
• arxiv keydann2021beyond

Related papers

• Square-root Regret Bounds For Continuous-time Episodic Markov Decision Processes (2022)2.26
• Optimism And Delays In Episodic Reinforcement Learning (2021)0.00
• Logarithmic Regret Of Exploration In Average Reward Markov Decision Processes (2025)0.00
• Provably Efficient Reinforcement Learning With Aggregated States (2019)0.00
• Sharp Variance-dependent Bounds In Reinforcement Learning: Best Of Both Worlds In Stochastic And Deterministic Environments (2023)0.00
• Dynamic Regret Of Online Markov Decision Processes (2022)0.00
• Regret Bounds For Discounted Mdps (2020)0.00
• Unifying PAC And Regret: Uniform PAC Bounds For Episodic Reinforcement Learning (2017)0.00