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

Download:

Current browse context:

cs.LG

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

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

Computer Science > Machine Learning

Title: Sublinear Optimal Policy Value Estimation in Contextual Bandits

Abstract: We study the problem of estimating the expected reward of the optimal policy in the stochastic disjoint linear bandit setting. We prove that for certain settings it is possible to obtain an accurate estimate of the optimal policy value even with a number of samples that is sublinear in the number that would be required to \emph{find} a policy that realizes a value close to this optima. We establish nearly matching information theoretic lower bounds, showing that our algorithm achieves near optimal estimation error. Finally, we demonstrate the effectiveness of our algorithm on joke recommendation and cancer inhibition dosage selection problems using real datasets.
Comments: Extended to the mixture of Gaussians setting
Subjects: Machine Learning (cs.LG); Machine Learning (stat.ML)
Cite as: arXiv:1912.06111 [cs.LG]
  (or arXiv:1912.06111v2 [cs.LG] for this version)

Submission history

From: Weihao Kong [view email]
[v1] Thu, 12 Dec 2019 18:20:11 GMT (180kb,D)
[v2] Fri, 13 Dec 2019 22:43:40 GMT (182kb,D)

Link back to: arXiv, form interface, contact.