Reinforcement Learning from Adversarial Preferences in Tabular MDPs

We introduce a new framework of episodic tabular Markov decision processes (MDPs) with adversarial preferences, which we refer to as preference-based MDPs (PbMDPs). Unlike standard episodic MDPs with adversarial losses, where the numerical value of the loss is directly observed, in PbMDPs the learner instead observes preferences between two candidate arms, which represent the choices being compared. In this work, we focus specifically on the setting where the reward functions are determined by Borda scores. We begin by establishing a regret lower bound for PbMDPs with Borda scores. As a preliminary step, we present a simple instance to prove a lower bound of Omega(√HSAT) for episodic MDPs with adversarial losses, where H is the number of steps per episode, S is the number of states, A is the number of actions, and T is the number of episodes. Leveraging this construction, we then derive a regret lower bound of Omega( (H2 S K)1/3 T2/3 ) for PbMDPs with Borda scores, where K is the number of arms. Next, we develop algorithms that achieve a regret bound of order T2/3. We first propose a global optimization approach based on online linear optimization over the set of all occupancy measures, achieving a regret bound of tildeO((H2 S2 K)1/3 T2/3 ) under known transitions. However, this approach suffers from suboptimal dependence on the potentially large number of states S and computational inefficiency. To address this, we propose a policy optimization algorithm whose regret is roughly bounded by tildeO( (H6 S K5)1/3 T2/3 ) under known transitions, and further extend the result to the unknown-transition setting.