Matches in SemOpenAlex for { <https://semopenalex.org/work/W3092042512> ?p ?o ?g. }
- W3092042512 abstract "In the classical multi-armed bandit problem, instance-dependent algorithms attain improved performance on easy problems with a gap between the best and second-best arm. Are similar guarantees possible for contextual bandits? While positive results are known for certain special cases, there is no general theory characterizing when and how instance-dependent regret bounds for contextual bandits can be achieved for rich, general classes of policies. We introduce a family of complexity measures that are both sufficient and necessary to obtain instance-dependent regret bounds. We then introduce new oracle-efficient algorithms which adapt to the gap whenever possible, while also attaining the minimax rate in the worst case. Finally, we provide structural results that tie together a number of complexity measures previously proposed throughout contextual bandits, reinforcement learning, and active learning and elucidate their role in determining the optimal instance-dependent regret. In a large-scale empirical evaluation, we find that our approach often gives superior results for challenging exploration problems. Turning our focus to reinforcement learning with function approximation, we develop new oracle-efficient algorithms for reinforcement learning with rich observations that obtain optimal gap-dependent sample complexity." @default.
- W3092042512 created "2020-10-15" @default.
- W3092042512 creator A5071385934 @default.
- W3092042512 creator A5075069827 @default.
- W3092042512 creator A5076656836 @default.
- W3092042512 creator A5078130712 @default.
- W3092042512 date "2020-10-07" @default.
- W3092042512 modified "2023-09-27" @default.
- W3092042512 title "Instance-Dependent Complexity of Contextual Bandits and Reinforcement Learning: A Disagreement-Based Perspective" @default.
- W3092042512 cites W1487320471 @default.
- W3092042512 cites W1506859583 @default.
- W3092042512 cites W1511694993 @default.
- W3092042512 cites W1595135966 @default.
- W3092042512 cites W1687382548 @default.
- W3092042512 cites W1983962754 @default.
- W3092042512 cites W1984876737 @default.
- W3092042512 cites W2009551863 @default.
- W3092042512 cites W2009772318 @default.
- W3092042512 cites W2029604420 @default.
- W3092042512 cites W2043919728 @default.
- W3092042512 cites W2049393399 @default.
- W3092042512 cites W2056138823 @default.
- W3092042512 cites W2057050711 @default.
- W3092042512 cites W2059529070 @default.
- W3092042512 cites W2066313192 @default.
- W3092042512 cites W2098258765 @default.
- W3092042512 cites W2100931509 @default.
- W3092042512 cites W2112420033 @default.
- W3092042512 cites W2114232233 @default.
- W3092042512 cites W2119738618 @default.
- W3092042512 cites W2119850747 @default.
- W3092042512 cites W2141514451 @default.
- W3092042512 cites W2145049547 @default.
- W3092042512 cites W2146502635 @default.
- W3092042512 cites W2154637324 @default.
- W3092042512 cites W2163840227 @default.
- W3092042512 cites W2166886563 @default.
- W3092042512 cites W2410087222 @default.
- W3092042512 cites W2545659366 @default.
- W3092042512 cites W2614208603 @default.
- W3092042512 cites W2734936460 @default.
- W3092042512 cites W2807644309 @default.
- W3092042512 cites W28787979 @default.
- W3092042512 cites W2890189120 @default.
- W3092042512 cites W2894130846 @default.
- W3092042512 cites W2912399346 @default.
- W3092042512 cites W2946218678 @default.
- W3092042512 cites W2962735187 @default.
- W3092042512 cites W2963402790 @default.
- W3092042512 cites W2963433763 @default.
- W3092042512 cites W2963434013 @default.
- W3092042512 cites W2963713569 @default.
- W3092042512 cites W2963938771 @default.
- W3092042512 cites W2963971282 @default.
- W3092042512 cites W2964054583 @default.
- W3092042512 cites W2964242659 @default.
- W3092042512 cites W2964278893 @default.
- W3092042512 cites W2964282972 @default.
- W3092042512 cites W2964284806 @default.
- W3092042512 cites W2964315724 @default.
- W3092042512 cites W2964341565 @default.
- W3092042512 cites W2971249033 @default.
- W3092042512 cites W2984671263 @default.
- W3092042512 cites W2990210896 @default.
- W3092042512 cites W2991929641 @default.
- W3092042512 cites W2995638039 @default.
- W3092042512 cites W3006052551 @default.
- W3092042512 cites W3011264688 @default.
- W3092042512 cites W3013725403 @default.
- W3092042512 cites W3026228308 @default.
- W3092042512 cites W3029753614 @default.
- W3092042512 cites W3034274776 @default.
- W3092042512 cites W3035810656 @default.
- W3092042512 cites W3037550606 @default.
- W3092042512 cites W3043226849 @default.
- W3092042512 cites W3045217160 @default.
- W3092042512 cites W3046395471 @default.
- W3092042512 cites W3100895096 @default.
- W3092042512 cites W3102366752 @default.
- W3092042512 cites W3124121179 @default.
- W3092042512 cites W50486269 @default.
- W3092042512 hasPublicationYear "2020" @default.
- W3092042512 type Work @default.
- W3092042512 sameAs 3092042512 @default.
- W3092042512 citedByCount "18" @default.
- W3092042512 countsByYear W30920425122021 @default.
- W3092042512 crossrefType "posted-content" @default.
- W3092042512 hasAuthorship W3092042512A5071385934 @default.
- W3092042512 hasAuthorship W3092042512A5075069827 @default.
- W3092042512 hasAuthorship W3092042512A5076656836 @default.
- W3092042512 hasAuthorship W3092042512A5078130712 @default.
- W3092042512 hasConcept C111472728 @default.
- W3092042512 hasConcept C115903868 @default.
- W3092042512 hasConcept C119857082 @default.
- W3092042512 hasConcept C126255220 @default.
- W3092042512 hasConcept C12713177 @default.
- W3092042512 hasConcept C138885662 @default.
- W3092042512 hasConcept C14036430 @default.
- W3092042512 hasConcept C149728462 @default.
- W3092042512 hasConcept C154945302 @default.