FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W588099057> ?p ?o ?g. }

• W588099057 abstract "Bandit games consist of single-state environments in which an agent must sequentiallychoose actions to take, for which rewards are given. The objective being to maximisethe cumulated reward, the agent naturally seeks to build a model of the relationshipbetween actions and rewards. The agent must both choose uncertain actions in orderto improve its model (exploration), and actions that are believed to yield high rewardsaccording to the model (exploitation). The choice of an action to take is called a playof an arm of the bandit, and the total number of plays may or may not be known inadvance.Algorithms designed to handle the exploration-exploitation dilemma were initiallymotivated by problems with rather small numbers of actions. But the ideas they werebased on have been extended to cases where the number of actions to choose from ismuch larger than the maximum possible number of plays. Several problems fall into thissetting, such as information retrieval with relevance feedback, where the system mustlearn what a user is looking for while serving relevant documents often enough, butalso global optimisation, where the search for an optimum is done by selecting whereto acquire potentially expensive samples of a target function. All have in common thesearch of large spaces.In this thesis, we focus on an algorithm based on the Gaussian Processes probabilisticmodel, often used in Bayesian optimisation, and the Upper Confidence Boundaction-selection heuristic that is popular in bandit algorithms. In addition to demonstratingthe advantages of the GP-UCB algorithm on an image retrieval problem, weshow how it can be adapted in order to search tree-structured spaces. We provide anefficient implementation, theoretical guarantees on the algorithm's performance, andempirical evidence that it handles large branching factors better than previous bandit-basedalgorithms, on synthetic trees." @default.
• W588099057 created "2016-06-24" @default.
• W588099057 creator A5068836983 @default.
• W588099057 date "2012-04-28" @default.
• W588099057 modified "2023-09-28" @default.
• W588099057 title "Bandit algorithms for searching large spaces" @default.
• W588099057 cites W1482780284 @default.
• W588099057 cites W1483202606 @default.
• W588099057 cites W1486950299 @default.
• W588099057 cites W1500868819 @default.
• W588099057 cites W1503422967 @default.
• W588099057 cites W1510073064 @default.
• W588099057 cites W1521084402 @default.
• W588099057 cites W1542148049 @default.
• W588099057 cites W1544496518 @default.
• W588099057 cites W1625390266 @default.
• W588099057 cites W1663973292 @default.
• W588099057 cites W1680189815 @default.
• W588099057 cites W1714211023 @default.
• W588099057 cites W1746819321 @default.
• W588099057 cites W1976355201 @default.
• W588099057 cites W1988360637 @default.
• W588099057 cites W1997840820 @default.
• W588099057 cites W2009551863 @default.
• W588099057 cites W2069462534 @default.
• W588099057 cites W2073384958 @default.
• W588099057 cites W2097329790 @default.
• W588099057 cites W2099201756 @default.
• W588099057 cites W2099679938 @default.
• W588099057 cites W2103581319 @default.
• W588099057 cites W2105066050 @default.
• W588099057 cites W2107386393 @default.
• W588099057 cites W2108114251 @default.
• W588099057 cites W2108738385 @default.
• W588099057 cites W2115519224 @default.
• W588099057 cites W2122853437 @default.
• W588099057 cites W2131627640 @default.
• W588099057 cites W2132350392 @default.
• W588099057 cites W2142971854 @default.
• W588099057 cites W2158807713 @default.
• W588099057 cites W2158858912 @default.
• W588099057 cites W2162979096 @default.
• W588099057 cites W2168405694 @default.
• W588099057 cites W2169511307 @default.
• W588099057 cites W2224879568 @default.
• W588099057 cites W2402456051 @default.
• W588099057 cites W2592746156 @default.
• W588099057 cites W2627076442 @default.
• W588099057 cites W2951665052 @default.
• W588099057 cites W2963110737 @default.
• W588099057 cites W2964172739 @default.
• W588099057 cites W50486269 @default.
• W588099057 cites W638603679 @default.
• W588099057 cites W83008820 @default.
• W588099057 hasPublicationYear "2012" @default.
• W588099057 type Work @default.
• W588099057 sameAs 588099057 @default.
• W588099057 citedByCount "0" @default.
• W588099057 crossrefType "dissertation" @default.
• W588099057 hasAuthorship W588099057A5068836983 @default.
• W588099057 hasConcept C119857082 @default.
• W588099057 hasConcept C121332964 @default.
• W588099057 hasConcept C126255220 @default.
• W588099057 hasConcept C14036430 @default.
• W588099057 hasConcept C154945302 @default.
• W588099057 hasConcept C158154518 @default.
• W588099057 hasConcept C166109690 @default.
• W588099057 hasConcept C169760540 @default.
• W588099057 hasConcept C173801870 @default.
• W588099057 hasConcept C17744445 @default.
• W588099057 hasConcept C199539241 @default.
• W588099057 hasConcept C26760741 @default.
• W588099057 hasConcept C2780791683 @default.
• W588099057 hasConcept C33923547 @default.
• W588099057 hasConcept C41008148 @default.
• W588099057 hasConcept C49937458 @default.
• W588099057 hasConcept C62520636 @default.
• W588099057 hasConcept C78458016 @default.
• W588099057 hasConcept C81917197 @default.
• W588099057 hasConcept C86803240 @default.
• W588099057 hasConceptScore W588099057C119857082 @default.
• W588099057 hasConceptScore W588099057C121332964 @default.
• W588099057 hasConceptScore W588099057C126255220 @default.
• W588099057 hasConceptScore W588099057C14036430 @default.
• W588099057 hasConceptScore W588099057C154945302 @default.
• W588099057 hasConceptScore W588099057C158154518 @default.
• W588099057 hasConceptScore W588099057C166109690 @default.
• W588099057 hasConceptScore W588099057C169760540 @default.
• W588099057 hasConceptScore W588099057C173801870 @default.
• W588099057 hasConceptScore W588099057C17744445 @default.
• W588099057 hasConceptScore W588099057C199539241 @default.
• W588099057 hasConceptScore W588099057C26760741 @default.
• W588099057 hasConceptScore W588099057C2780791683 @default.
• W588099057 hasConceptScore W588099057C33923547 @default.
• W588099057 hasConceptScore W588099057C41008148 @default.
• W588099057 hasConceptScore W588099057C49937458 @default.
• W588099057 hasConceptScore W588099057C62520636 @default.
• W588099057 hasConceptScore W588099057C78458016 @default.
• W588099057 hasConceptScore W588099057C81917197 @default.
• W588099057 hasConceptScore W588099057C86803240 @default.

• next