We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations

DBLP - CS Bibliography


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: Submodular Bandit Problem Under Multiple Constraints

Abstract: The linear submodular bandit problem was proposed to simultaneously address diversified retrieval and online learning in a recommender system. If there is no uncertainty, this problem is equivalent to a submodular maximization problem under a cardinality constraint. However, in some situations, recommendation lists should satisfy additional constraints such as budget constraints, other than a cardinality constraint. Thus, motivated by diversified retrieval considering budget constraints, we introduce a submodular bandit problem under the intersection of $l$ knapsacks and a $k$-system constraint. Here $k$-system constraints form a very general class of constraints including cardinality constraints and the intersection of $k$ matroid constraints. To solve this problem, we propose a non-greedy algorithm that adaptively focuses on a standard or modified upper-confidence bound. We provide a high-probability upper bound of an approximation regret, where the approximation ratio matches that of a fast offline algorithm. Moreover, we perform experiments under various combinations of constraints using a synthetic and two real-world datasets and demonstrate that our proposed methods outperform the existing baselines.
Comments: accepted at UAI 2020, minor mistakes fixed
Subjects: Machine Learning (cs.LG); Machine Learning (stat.ML)
Cite as: arXiv:2006.00661 [cs.LG]
  (or arXiv:2006.00661v5 [cs.LG] for this version)

Submission history

From: Sho Takemori Ph.D [view email]
[v1] Mon, 1 Jun 2020 01:28:44 GMT (599kb,AD)
[v2] Wed, 3 Jun 2020 06:59:23 GMT (599kb,AD)
[v3] Fri, 31 Jul 2020 04:10:35 GMT (599kb,AD)
[v4] Mon, 26 Oct 2020 05:12:46 GMT (599kb,AD)
[v5] Mon, 29 Mar 2021 02:02:19 GMT (599kb,AD)

Link back to: arXiv, form interface, contact.