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

About

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.

Masrour Zoghi, Shimon Whiteson, Remi Munos, Maarten de Rijke · 2013

Related benchmarks

Task	Dataset	Result
Multi-objective preference selection	OpenAI models Setup A3: K=9, M*=2	Cost ($)0.117	8
Multi-objective preference selection	OpenAI models Setup A4: K=5, M*=5	Cost ($)0.118	8
Multi-objective preference selection	OpenAI models Setup A1: K=5, M*=2	Cost ($)0.161	8
Multi-objective preference selection	OpenAI models Setup A2: K=7, M*=10	Cost ($)0.116	8