FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3169119266> ?p ?o ?g. }

• W3169119266 abstract "Estimating the gradients for binary variables is a task that arises frequently in various domains, such as training discrete latent variable models. What has been commonly used is a REINFORCE based Monte Carlo estimation method that uses either independent samples or pairs of negatively correlated samples. To better utilize more than two samples, we propose ARMS, an Antithetic REINFORCE-based Multi-Sample gradient estimator. ARMS uses a copula to generate any number of mutually antithetic samples. It is unbiased, has low variance, and generalizes both DisARM, which we show to be ARMS with two samples, and the leave-one-out REINFORCE (LOORF) estimator, which is ARMS with uncorrelated samples. We evaluate ARMS on several datasets for training generative models, and our experimental results show that it outperforms competing methods. We also develop a version of ARMS for optimizing the multi-sample variational bound, and show that it outperforms both VIMCO and DisARM. The code is publicly available." @default.
• W3169119266 created "2021-06-22" @default.
• W3169119266 creator A5010394308 @default.
• W3169119266 creator A5014918622 @default.
• W3169119266 date "2021-05-28" @default.
• W3169119266 modified "2023-09-27" @default.
• W3169119266 title "ARMS: Antithetic-REINFORCE-Multi-Sample Gradient for Binary Variables" @default.
• W3169119266 cites W1516111018 @default.
• W3169119266 cites W1959608418 @default.
• W3169119266 cites W2001249527 @default.
• W3169119266 cites W2046765929 @default.
• W3169119266 cites W2119196781 @default.
• W3169119266 cites W2119717200 @default.
• W3169119266 cites W2122262818 @default.
• W3169119266 cites W2187669537 @default.
• W3169119266 cites W2238910925 @default.
• W3169119266 cites W2242818861 @default.
• W3169119266 cites W2547875792 @default.
• W3169119266 cites W2548228487 @default.
• W3169119266 cites W2602076750 @default.
• W3169119266 cites W2803107756 @default.
• W3169119266 cites W2909637611 @default.
• W3169119266 cites W2946426880 @default.
• W3169119266 cites W2962839265 @default.
• W3169119266 cites W2962897886 @default.
• W3169119266 cites W2962994101 @default.
• W3169119266 cites W2963173382 @default.
• W3169119266 cites W2963275229 @default.
• W3169119266 cites W2963351648 @default.
• W3169119266 cites W2963675378 @default.
• W3169119266 cites W2963851840 @default.
• W3169119266 cites W2963936947 @default.
• W3169119266 cites W2964121744 @default.
• W3169119266 cites W3035314628 @default.
• W3169119266 cites W3082278101 @default.
• W3169119266 cites W3099014706 @default.
• W3169119266 cites W584423923 @default.
• W3169119266 hasPublicationYear "2021" @default.
• W3169119266 type Work @default.
• W3169119266 sameAs 3169119266 @default.
• W3169119266 citedByCount "0" @default.
• W3169119266 crossrefType "posted-content" @default.
• W3169119266 hasAuthorship W3169119266A5010394308 @default.
• W3169119266 hasAuthorship W3169119266A5014918622 @default.
• W3169119266 hasConcept C105795698 @default.
• W3169119266 hasConcept C11413529 @default.
• W3169119266 hasConcept C121955636 @default.
• W3169119266 hasConcept C134306372 @default.
• W3169119266 hasConcept C144133560 @default.
• W3169119266 hasConcept C149782125 @default.
• W3169119266 hasConcept C169345407 @default.
• W3169119266 hasConcept C17618745 @default.
• W3169119266 hasConcept C182365436 @default.
• W3169119266 hasConcept C185429906 @default.
• W3169119266 hasConcept C185592680 @default.
• W3169119266 hasConcept C19499675 @default.
• W3169119266 hasConcept C196083921 @default.
• W3169119266 hasConcept C198531522 @default.
• W3169119266 hasConcept C33923547 @default.
• W3169119266 hasConcept C41008148 @default.
• W3169119266 hasConcept C43617362 @default.
• W3169119266 hasConcept C48372109 @default.
• W3169119266 hasConcept C51167844 @default.
• W3169119266 hasConcept C94375191 @default.
• W3169119266 hasConceptScore W3169119266C105795698 @default.
• W3169119266 hasConceptScore W3169119266C11413529 @default.
• W3169119266 hasConceptScore W3169119266C121955636 @default.
• W3169119266 hasConceptScore W3169119266C134306372 @default.
• W3169119266 hasConceptScore W3169119266C144133560 @default.
• W3169119266 hasConceptScore W3169119266C149782125 @default.
• W3169119266 hasConceptScore W3169119266C169345407 @default.
• W3169119266 hasConceptScore W3169119266C17618745 @default.
• W3169119266 hasConceptScore W3169119266C182365436 @default.
• W3169119266 hasConceptScore W3169119266C185429906 @default.
• W3169119266 hasConceptScore W3169119266C185592680 @default.
• W3169119266 hasConceptScore W3169119266C19499675 @default.
• W3169119266 hasConceptScore W3169119266C196083921 @default.
• W3169119266 hasConceptScore W3169119266C198531522 @default.
• W3169119266 hasConceptScore W3169119266C33923547 @default.
• W3169119266 hasConceptScore W3169119266C41008148 @default.
• W3169119266 hasConceptScore W3169119266C43617362 @default.
• W3169119266 hasConceptScore W3169119266C48372109 @default.
• W3169119266 hasConceptScore W3169119266C51167844 @default.
• W3169119266 hasConceptScore W3169119266C94375191 @default.
• W3169119266 hasLocation W31691192661 @default.
• W3169119266 hasOpenAccess W3169119266 @default.
• W3169119266 hasPrimaryLocation W31691192661 @default.
• W3169119266 hasRelatedWork W1685295223 @default.
• W3169119266 hasRelatedWork W2399445579 @default.
• W3169119266 hasRelatedWork W2470396004 @default.
• W3169119266 hasRelatedWork W2594399450 @default.
• W3169119266 hasRelatedWork W2623267204 @default.
• W3169119266 hasRelatedWork W2781952440 @default.
• W3169119266 hasRelatedWork W2785763383 @default.
• W3169119266 hasRelatedWork W2809763245 @default.
• W3169119266 hasRelatedWork W2971686979 @default.
• W3169119266 hasRelatedWork W2978075675 @default.
• W3169119266 hasRelatedWork W2990013650 @default.
• W3169119266 hasRelatedWork W2995657162 @default.
• W3169119266 hasRelatedWork W2996810036 @default.

• next