Università degli Studi dell'Insubria Insubria Space

InsubriaSPACE - Thesis PhD Repository >
Insubria Thesis Repository >
01 - Tesi di dottorato >

Please use this identifier to cite or link to this item: http://hdl.handle.net/10277/729

Authors: Li, Shuai
Internal Tutor: GENTILE, CLAUDIO
Title: The art of clustering bandits.
Abstract: Multi-armed bandit problems are receiving a great deal of attention because they adequately formalize the exploration-exploitation trade-offs arising in several industrially relevant applications, such as online advertisement and, more generally, recommendation systems. In many cases, however, these applications have a strong social component, whose integration in the bandit algorithms could lead to a dramatic performance increase. For instance, we may want to serve content to a group of users by taking advantage of an underlying network of social relationships among them. The purpose of this thesis is to introduce novel and principled algorithmic approaches to the solution of such networked bandit problems. Starting from a global (Laplacian-based) strategy which allocates a bandit algorithm to each network node (user), and allows it to "share" signals (contexts and payoffs) with the neghboring nodes, our goal is to derive and experimentally test more scalable approaches based on different ways of clustering the graph nodes. More importantly, we shall investigate the case when the graph structure is not given ahead of time, and has to be inferred based on past user behavior. A general difficulty arising in such practical scenarios is that data sequences are typically nonstationary, implying that traditional statistical inference methods should be used cautiously, possibly replacing them with by more robust nonstochastic (e.g., game-theoretic) inference methods. In this thesis, we will firstly introduce the centralized clustering bandits. Then, we propose the corresponding solution in decentralized scenario. After that, we explain the generic collaborative clustering bandits. Finally, we extend and showcase the state-of-the-art clustering bandits that we developed in the quantification problem.
Keywords: Information retrieval, data mining, machine learning
Issue Date: 2016
Language: eng
Doctoral course: Informatica e matematica del calcolo
Academic cycle: 29
Publisher: Università degli Studi dell'Insubria
Citation: Li, S.The art of clustering bandits. (Doctoral Thesis, Università degli Studi dell'Insubria, 2016).

Files in This Item:

File Description SizeFormatVisibility
PhD_Thesis_Lishuai_completa.pdftesto completo tesi2,93 MBAdobe PDFView/Open

This item is licensed under a Creative Commons License
Creative Commons

Items in InsubriaSPACE are protected by copyright, with all rights reserved, unless otherwise indicated.

Share this record




Stumble it!



  ICT Support, development & maintenance are provided by the AePIC team @ CILEA. Powered on DSpace Software.  Feedback