Timezone: »

 
Poster
An Empirical Process Approach to the Union Bound: Practical Algorithms for Combinatorial and Linear Bandits
Julian Katz-Samuels · Lalit Jain · zohar karnin · Kevin Jamieson

Wed Dec 09 09:00 PM -- 11:00 PM (PST) @ Poster Session 4 #1282
in Poster Session 5 »

This paper proposes near-optimal algorithms for the pure-exploration linear bandit problem in the fixed confidence and fixed budget settings. Leveraging ideas from the theory of suprema of empirical processes, we provide an algorithm whose sample complexity scales with the geometry of the instance and avoids an explicit union bound over the number of arms. Unlike previous approaches which sample based on minimizing a worst-case variance (e.g. G-optimal design), we define an experimental design objective based on the Gaussian-width of the underlying arm set. We provide a novel lower bound in terms of this objective that highlights its fundamental role in the sample complexity. The sample complexity of our fixed confidence algorithm matches this lower bound, and in addition is computationally efficient for combinatorial classes, e.g. shortest-path, matchings and matroids, where the arm sets can be exponentially large in the dimension. Finally, we propose the first algorithm for linear bandits in the the fixed budget setting. Its guarantee matches our lower bound up to logarithmic factors.

Author Information

Julian Katz-Samuels (University of Washington)
Lalit Jain (University of Washington)
zohar karnin (Amazon)
Kevin Jamieson (U Washington)

More from the Same Authors