Relative Upper Confidence Bound for the K-Armed Dueling Bandit Problem

Abstract

This paper proposes a new method for the K-armed dueling bandit problem, a variation on the regular K-armed bandit problem that offers only relative feedback about pairs of arms. Our approach extends the Upper Confidence Bound algorithm to the relative setting by using estimates of the pairwise probabilities to select a promising arm and applying Upper Confidence Bound with the winner as a benchmark. We prove a finite-time regret bound of order O(log t). In addition, our empirical results using real data from an information retrieval application show that it greatly outperforms the state of the art.

13 pages, 6 figures

References in corpus (3)

• Stochastic simultaneous optimistic optimization
• Exponential Regret Bounds for Gaussian Process Bandits with Deterministic Observations
• An Adaptive Algorithm for Finite Stochastic Partial Monitoring

Cited by in corpus (15)

• Preference-based Online Learning with Dueling Bandits: A Survey
• Copeland Dueling Bandits
• Copeland Dueling Bandit Problem: Regret Lower Bound, Optimal Algorithm, and Computationally Efficient Algorithm
• Combinatorial Pure Exploration of Dueling Bandit
• MergeDTS: A Method for Effective Large-Scale Online Ranker Evaluation
• Dueling RL: Reinforcement Learning with Trajectory Preferences
• Dueling Bandits With Weak Regret
• Regret Minimization in Stochastic Contextual Dueling Bandits
• Preferential Batch Bayesian Optimization
• Decoy Bandits Dueling on a Poset
• Best-item Learning in Random Utility Models with Subset Choices
• Adversarial Dueling Bandits
• Choice functions based multi-objective Bayesian optimisation
• Dueling Bandits with Adversarial Sleeping
• Efficient and Optimal Algorithms for Contextual Dueling Bandits under Realizability