Matches in SemOpenAlex for { <https://semopenalex.org/work/W2948598159> ?p ?o ?g. }
- W2948598159 abstract "Estimation of importance sampling weights for off-policy evaluation of contextual bandits often results in imbalance - a mismatch between the desired and the actual distribution of state-action pairs after weighting. In this work we present balanced off-policy evaluation (B-OPE), a generic method for estimating weights which minimize this imbalance. Estimation of these weights reduces to a binary classification problem regardless of action type. We show that minimizing the risk of the classifier implies minimization of imbalance to the desired counterfactual distribution of state-action pairs. The classifier loss is tied to the error of the off-policy estimate, allowing for easy tuning of hyperparameters. We provide experimental evidence that B-OPE improves weighting-based approaches for offline policy evaluation in both discrete and continuous action spaces." @default.
- W2948598159 created "2019-06-14" @default.
- W2948598159 creator A5063445465 @default.
- W2948598159 creator A5063845534 @default.
- W2948598159 creator A5074165520 @default.
- W2948598159 date "2019-06-09" @default.
- W2948598159 modified "2023-10-01" @default.
- W2948598159 title "Balanced off-policy evaluation in general action spaces" @default.
- W2948598159 cites W123476658 @default.
- W2948598159 cites W1532264762 @default.
- W2948598159 cites W1809653203 @default.
- W2948598159 cites W1966026565 @default.
- W2948598159 cites W1999188374 @default.
- W2948598159 cites W2033468335 @default.
- W2948598159 cites W2042179218 @default.
- W2948598159 cites W2044116626 @default.
- W2948598159 cites W2062947384 @default.
- W2948598159 cites W2064769840 @default.
- W2948598159 cites W2101557761 @default.
- W2948598159 cites W2102689555 @default.
- W2948598159 cites W2103459159 @default.
- W2948598159 cites W2112483442 @default.
- W2948598159 cites W2120817734 @default.
- W2948598159 cites W2121506959 @default.
- W2948598159 cites W2129606454 @default.
- W2948598159 cites W2132324013 @default.
- W2948598159 cites W2137370054 @default.
- W2948598159 cites W2138909795 @default.
- W2948598159 cites W2139529156 @default.
- W2948598159 cites W2150291618 @default.
- W2948598159 cites W2170612786 @default.
- W2948598159 cites W2212660284 @default.
- W2948598159 cites W2439299270 @default.
- W2948598159 cites W2519411794 @default.
- W2948598159 cites W2585690194 @default.
- W2948598159 cites W2734936460 @default.
- W2948598159 cites W2787436538 @default.
- W2948598159 cites W2811380766 @default.
- W2948598159 cites W2949950578 @default.
- W2948598159 cites W2962736281 @default.
- W2948598159 cites W2962785510 @default.
- W2948598159 cites W2963323139 @default.
- W2948598159 cites W2963448230 @default.
- W2948598159 cites W2964000438 @default.
- W2948598159 cites W2964297722 @default.
- W2948598159 cites W2979978950 @default.
- W2948598159 cites W3098679278 @default.
- W2948598159 cites W3122812581 @default.
- W2948598159 cites W3125357717 @default.
- W2948598159 cites W362526619 @default.
- W2948598159 cites W91088564 @default.
- W2948598159 hasPublicationYear "2019" @default.
- W2948598159 type Work @default.
- W2948598159 sameAs 2948598159 @default.
- W2948598159 citedByCount "1" @default.
- W2948598159 countsByYear W29485981592019 @default.
- W2948598159 crossrefType "posted-content" @default.
- W2948598159 hasAuthorship W2948598159A5063445465 @default.
- W2948598159 hasAuthorship W2948598159A5063845534 @default.
- W2948598159 hasAuthorship W2948598159A5074165520 @default.
- W2948598159 hasConcept C108650721 @default.
- W2948598159 hasConcept C11413529 @default.
- W2948598159 hasConcept C121332964 @default.
- W2948598159 hasConcept C12267149 @default.
- W2948598159 hasConcept C126255220 @default.
- W2948598159 hasConcept C126838900 @default.
- W2948598159 hasConcept C147764199 @default.
- W2948598159 hasConcept C149782125 @default.
- W2948598159 hasConcept C154945302 @default.
- W2948598159 hasConcept C15744967 @default.
- W2948598159 hasConcept C183115368 @default.
- W2948598159 hasConcept C2780791683 @default.
- W2948598159 hasConcept C33923547 @default.
- W2948598159 hasConcept C41008148 @default.
- W2948598159 hasConcept C48372109 @default.
- W2948598159 hasConcept C62520636 @default.
- W2948598159 hasConcept C66905080 @default.
- W2948598159 hasConcept C71924100 @default.
- W2948598159 hasConcept C77805123 @default.
- W2948598159 hasConcept C8642999 @default.
- W2948598159 hasConcept C94375191 @default.
- W2948598159 hasConcept C95623464 @default.
- W2948598159 hasConceptScore W2948598159C108650721 @default.
- W2948598159 hasConceptScore W2948598159C11413529 @default.
- W2948598159 hasConceptScore W2948598159C121332964 @default.
- W2948598159 hasConceptScore W2948598159C12267149 @default.
- W2948598159 hasConceptScore W2948598159C126255220 @default.
- W2948598159 hasConceptScore W2948598159C126838900 @default.
- W2948598159 hasConceptScore W2948598159C147764199 @default.
- W2948598159 hasConceptScore W2948598159C149782125 @default.
- W2948598159 hasConceptScore W2948598159C154945302 @default.
- W2948598159 hasConceptScore W2948598159C15744967 @default.
- W2948598159 hasConceptScore W2948598159C183115368 @default.
- W2948598159 hasConceptScore W2948598159C2780791683 @default.
- W2948598159 hasConceptScore W2948598159C33923547 @default.
- W2948598159 hasConceptScore W2948598159C41008148 @default.
- W2948598159 hasConceptScore W2948598159C48372109 @default.
- W2948598159 hasConceptScore W2948598159C62520636 @default.
- W2948598159 hasConceptScore W2948598159C66905080 @default.
- W2948598159 hasConceptScore W2948598159C71924100 @default.