Matches in SemOpenAlex for { <https://semopenalex.org/work/W3028242018> ?p ?o ?g. }
- W3028242018 abstract "Greedy-GQ is an off-policy two timescale algorithm for optimal control in reinforcement learning. This paper develops the first finite-sample analysis for the Greedy-GQ algorithm with linear function approximation under Markovian noise. Our finite-sample analysis provides theoretical justification for choosing stepsizes for this two timescale algorithm for faster convergence in practice, and suggests a trade-off between the convergence rate and the quality of the obtained policy. Our paper extends the finite-sample analyses of two timescale reinforcement learning algorithms from policy evaluation to optimal control, which is of more practical interest. Specifically, in contrast to existing finite-sample analyses for two timescale methods, e.g., GTD, GTD2 and TDC, where their objective functions are convex, the objective function of the Greedy-GQ algorithm is non-convex. Moreover, the Greedy-GQ algorithm is also not a linear two-timescale stochastic approximation algorithm. Our techniques in this paper provide a general framework for finite-sample analysis of non-convex value-based reinforcement learning algorithms for optimal control." @default.
- W3028242018 created "2020-05-29" @default.
- W3028242018 creator A5012545205 @default.
- W3028242018 creator A5042736792 @default.
- W3028242018 date "2020-05-20" @default.
- W3028242018 modified "2023-09-27" @default.
- W3028242018 title "Finite-sample Analysis of Greedy-GQ with Linear Function Approximation under Markovian Noise" @default.
- W3028242018 cites W13294968 @default.
- W3028242018 cites W1646707810 @default.
- W3028242018 cites W1977655452 @default.
- W3028242018 cites W2075268401 @default.
- W3028242018 cites W2098432798 @default.
- W3028242018 cites W2105336508 @default.
- W3028242018 cites W2121863487 @default.
- W3028242018 cites W2151661095 @default.
- W3028242018 cites W2257979135 @default.
- W3028242018 cites W2395162158 @default.
- W3028242018 cites W2473364827 @default.
- W3028242018 cites W2626384641 @default.
- W3028242018 cites W2776728688 @default.
- W3028242018 cites W2798888671 @default.
- W3028242018 cites W2950882283 @default.
- W3028242018 cites W2964043796 @default.
- W3028242018 cites W2964123095 @default.
- W3028242018 cites W2964203948 @default.
- W3028242018 cites W2970128053 @default.
- W3028242018 cites W2970216099 @default.
- W3028242018 cites W2970961807 @default.
- W3028242018 cites W2970999177 @default.
- W3028242018 cites W3038797447 @default.
- W3028242018 cites W3136903997 @default.
- W3028242018 cites W3139377883 @default.
- W3028242018 hasPublicationYear "2020" @default.
- W3028242018 type Work @default.
- W3028242018 sameAs 3028242018 @default.
- W3028242018 citedByCount "5" @default.
- W3028242018 countsByYear W30282420182020 @default.
- W3028242018 countsByYear W30282420182021 @default.
- W3028242018 crossrefType "posted-content" @default.
- W3028242018 hasAuthorship W3028242018A5012545205 @default.
- W3028242018 hasAuthorship W3028242018A5042736792 @default.
- W3028242018 hasConcept C105795698 @default.
- W3028242018 hasConcept C106189395 @default.
- W3028242018 hasConcept C11413529 @default.
- W3028242018 hasConcept C121332964 @default.
- W3028242018 hasConcept C126255220 @default.
- W3028242018 hasConcept C127162648 @default.
- W3028242018 hasConcept C14036430 @default.
- W3028242018 hasConcept C14646407 @default.
- W3028242018 hasConcept C154945302 @default.
- W3028242018 hasConcept C159886148 @default.
- W3028242018 hasConcept C162324750 @default.
- W3028242018 hasConcept C170131372 @default.
- W3028242018 hasConcept C188116033 @default.
- W3028242018 hasConcept C198531522 @default.
- W3028242018 hasConcept C26517878 @default.
- W3028242018 hasConcept C2777303404 @default.
- W3028242018 hasConcept C31258907 @default.
- W3028242018 hasConcept C33923547 @default.
- W3028242018 hasConcept C38652104 @default.
- W3028242018 hasConcept C41008148 @default.
- W3028242018 hasConcept C50522688 @default.
- W3028242018 hasConcept C51823790 @default.
- W3028242018 hasConcept C55479107 @default.
- W3028242018 hasConcept C57869625 @default.
- W3028242018 hasConcept C78458016 @default.
- W3028242018 hasConcept C86803240 @default.
- W3028242018 hasConcept C91575142 @default.
- W3028242018 hasConcept C97355855 @default.
- W3028242018 hasConcept C97541855 @default.
- W3028242018 hasConceptScore W3028242018C105795698 @default.
- W3028242018 hasConceptScore W3028242018C106189395 @default.
- W3028242018 hasConceptScore W3028242018C11413529 @default.
- W3028242018 hasConceptScore W3028242018C121332964 @default.
- W3028242018 hasConceptScore W3028242018C126255220 @default.
- W3028242018 hasConceptScore W3028242018C127162648 @default.
- W3028242018 hasConceptScore W3028242018C14036430 @default.
- W3028242018 hasConceptScore W3028242018C14646407 @default.
- W3028242018 hasConceptScore W3028242018C154945302 @default.
- W3028242018 hasConceptScore W3028242018C159886148 @default.
- W3028242018 hasConceptScore W3028242018C162324750 @default.
- W3028242018 hasConceptScore W3028242018C170131372 @default.
- W3028242018 hasConceptScore W3028242018C188116033 @default.
- W3028242018 hasConceptScore W3028242018C198531522 @default.
- W3028242018 hasConceptScore W3028242018C26517878 @default.
- W3028242018 hasConceptScore W3028242018C2777303404 @default.
- W3028242018 hasConceptScore W3028242018C31258907 @default.
- W3028242018 hasConceptScore W3028242018C33923547 @default.
- W3028242018 hasConceptScore W3028242018C38652104 @default.
- W3028242018 hasConceptScore W3028242018C41008148 @default.
- W3028242018 hasConceptScore W3028242018C50522688 @default.
- W3028242018 hasConceptScore W3028242018C51823790 @default.
- W3028242018 hasConceptScore W3028242018C55479107 @default.
- W3028242018 hasConceptScore W3028242018C57869625 @default.
- W3028242018 hasConceptScore W3028242018C78458016 @default.
- W3028242018 hasConceptScore W3028242018C86803240 @default.
- W3028242018 hasConceptScore W3028242018C91575142 @default.
- W3028242018 hasConceptScore W3028242018C97355855 @default.
- W3028242018 hasConceptScore W3028242018C97541855 @default.
- W3028242018 hasLocation W30282420181 @default.