In this thesis we address the multi-armed bandit (MAB) problem with stochastic rewards and correlated arms. Particularly, we investigate the case when the expected rewards are a Lipschitz function of the arm, and the learning to rank problem, as viewed from a MAB perspective. For the former, we derive a problem specific lower bound and propose both an asymptotically optimal algorithm (OSLB) and a (pareto)optimal, algorithm (POSLB). For the latter, we construct the regret lower bound and determine its closed form for some particular settings, as well as propose two asymptotically optimal algorithms PIE and PIE-C. For all algorithms mentioned above, we present performance analysis in the form of theoretical regret guarantees as well as numerical evaluation on artificial datasets as well as real-world datasets, in the case of PIE and PIE-C. / <p>QC 20160223</p>
| Identifer | oai:union.ndltd.org:UPSALLA1/oai:DiVA.org:kth-182816 |
| Date | January 2016 |
| Creators | Magureanu, Stefan |
| Publisher | KTH, Reglerteknik, Stockholm |
| Source Sets | DiVA Archive at Upsalla University |
| Language | English |
| Detected Language | English |
| Type | Licentiate thesis, monograph, info:eu-repo/semantics/masterThesis, text |
| Format | application/pdf |
| Rights | info:eu-repo/semantics/openAccess |
| Relation | TRITA-EE, 1653-5146 ; 2016:021 |
Page generated in 0.0018 seconds