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: Top-K Ranking Deep Contextual Bandits for Information Selection Systems

Abstract: In today's technology environment, information is abundant, dynamic, and heterogeneous in nature. Automated filtering and prioritization of information is based on the distinction between whether the information adds substantial value toward one's goal or not. Contextual multi-armed bandit has been widely used for learning to filter contents and prioritize according to user interest or relevance. Learn-to-Rank technique optimizes the relevance ranking on items, allowing the contents to be selected accordingly. We propose a novel approach to top-K rankings under the contextual multi-armed bandit framework. We model the stochastic reward function with a neural network to allow non-linear approximation to learn the relationship between rewards and contexts. We demonstrate the approach and evaluate the the performance of learning from the experiments using real world data sets in simulated scenarios. Empirical results show that this approach performs well under the complexity of a reward structure and high dimensional contextual features.
Subjects: Machine Learning (cs.LG); Information Retrieval (cs.IR)
Journal reference: 2021 IEEE International Conference on Systems, Man, and Cybernetics (SMC), 2021, pp. 2209-2214
DOI: 10.1109/SMC52423.2021.9658912
Cite as: arXiv:2201.13287 [cs.LG]
  (or arXiv:2201.13287v1 [cs.LG] for this version)

Submission history

From: Michael Rawson [view email]
[v1] Fri, 28 Jan 2022 15:10:44 GMT (2056kb,D)

Link back to: arXiv, form interface, contact.