FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3048796270> ?p ?o ?g. }

• W3048796270 abstract "Replica exchange Monte Carlo (reMC), also known as parallel tempering, is an important technique for accelerating the convergence of the conventional Markov Chain Monte Carlo (MCMC) algorithms. However, such a method requires the evaluation of the energy function based on the full dataset and is not scalable to big data. The naive implementation of reMC in mini-batch settings introduces large biases, which cannot be directly extended to the stochastic gradient MCMC (SGMCMC), the standard sampling method for simulating from deep neural networks (DNNs). In this paper, we propose an adaptive replica exchange SGMCMC (reSGMCMC) to automatically correct the bias and study the corresponding properties. The analysis implies an acceleration-accuracy trade-off in the numerical discretization of a Markov jump process in a stochastic environment. Empirically, we test the algorithm through extensive experiments on various setups and obtain the state-of-the-art results on CIFAR10, CIFAR100, and SVHN in both supervised learning and semi-supervised learning tasks." @default.
• W3048796270 created "2020-08-18" @default.
• W3048796270 creator A5011441236 @default.
• W3048796270 creator A5043652906 @default.
• W3048796270 creator A5076740334 @default.
• W3048796270 creator A5078138445 @default.
• W3048796270 creator A5085287370 @default.
• W3048796270 date "2020-08-12" @default.
• W3048796270 modified "2023-09-29" @default.
• W3048796270 title "Non-convex Learning via Replica Exchange Stochastic Gradient MCMC" @default.
• W3048796270 cites W104454400 @default.
• W3048796270 cites W123701566 @default.
• W3048796270 cites W1534145457 @default.
• W3048796270 cites W1546689893 @default.
• W3048796270 cites W1579781813 @default.
• W3048796270 cites W1591798773 @default.
• W3048796270 cites W1672071289 @default.
• W3048796270 cites W1972094069 @default.
• W3048796270 cites W1973575289 @default.
• W3048796270 cites W1983628095 @default.
• W3048796270 cites W1994616650 @default.
• W3048796270 cites W2000412346 @default.
• W3048796270 cites W2003268963 @default.
• W3048796270 cites W2013145865 @default.
• W3048796270 cites W2024060531 @default.
• W3048796270 cites W2070005565 @default.
• W3048796270 cites W2097091525 @default.
• W3048796270 cites W2115067168 @default.
• W3048796270 cites W2126750122 @default.
• W3048796270 cites W2128709328 @default.
• W3048796270 cites W2142859603 @default.
• W3048796270 cites W2144193737 @default.
• W3048796270 cites W2157327112 @default.
• W3048796270 cites W2159902196 @default.
• W3048796270 cites W2165652770 @default.
• W3048796270 cites W2167433878 @default.
• W3048796270 cites W2171032649 @default.
• W3048796270 cites W2186210550 @default.
• W3048796270 cites W2238987678 @default.
• W3048796270 cites W2594030462 @default.
• W3048796270 cites W2769986856 @default.
• W3048796270 cites W2808319839 @default.
• W3048796270 cites W2909262725 @default.
• W3048796270 cites W2962732346 @default.
• W3048796270 cites W2962839956 @default.
• W3048796270 cites W2962932339 @default.
• W3048796270 cites W2963050127 @default.
• W3048796270 cites W2963481221 @default.
• W3048796270 cites W2963874210 @default.
• W3048796270 cites W2963942415 @default.
• W3048796270 cites W2964111771 @default.
• W3048796270 cites W2970190082 @default.
• W3048796270 cites W2971225791 @default.
• W3048796270 cites W3035152412 @default.
• W3048796270 cites W423603040 @default.
• W3048796270 hasPublicationYear "2020" @default.
• W3048796270 type Work @default.
• W3048796270 sameAs 3048796270 @default.
• W3048796270 citedByCount "1" @default.
• W3048796270 countsByYear W30487962702020 @default.
• W3048796270 crossrefType "posted-content" @default.
• W3048796270 hasAuthorship W3048796270A5011441236 @default.
• W3048796270 hasAuthorship W3048796270A5043652906 @default.
• W3048796270 hasAuthorship W3048796270A5076740334 @default.
• W3048796270 hasAuthorship W3048796270A5078138445 @default.
• W3048796270 hasAuthorship W3048796270A5085287370 @default.
• W3048796270 hasConcept C105795698 @default.
• W3048796270 hasConcept C107673813 @default.
• W3048796270 hasConcept C111350023 @default.
• W3048796270 hasConcept C11413529 @default.
• W3048796270 hasConcept C119857082 @default.
• W3048796270 hasConcept C126255220 @default.
• W3048796270 hasConcept C13153151 @default.
• W3048796270 hasConcept C142362112 @default.
• W3048796270 hasConcept C153349607 @default.
• W3048796270 hasConcept C154945302 @default.
• W3048796270 hasConcept C162324750 @default.
• W3048796270 hasConcept C187653413 @default.
• W3048796270 hasConcept C19499675 @default.
• W3048796270 hasConcept C2775937380 @default.
• W3048796270 hasConcept C2777303404 @default.
• W3048796270 hasConcept C33923547 @default.
• W3048796270 hasConcept C41008148 @default.
• W3048796270 hasConcept C48044578 @default.
• W3048796270 hasConcept C50522688 @default.
• W3048796270 hasConcept C77088390 @default.
• W3048796270 hasConceptScore W3048796270C105795698 @default.
• W3048796270 hasConceptScore W3048796270C107673813 @default.
• W3048796270 hasConceptScore W3048796270C111350023 @default.
• W3048796270 hasConceptScore W3048796270C11413529 @default.
• W3048796270 hasConceptScore W3048796270C119857082 @default.
• W3048796270 hasConceptScore W3048796270C126255220 @default.
• W3048796270 hasConceptScore W3048796270C13153151 @default.
• W3048796270 hasConceptScore W3048796270C142362112 @default.
• W3048796270 hasConceptScore W3048796270C153349607 @default.
• W3048796270 hasConceptScore W3048796270C154945302 @default.
• W3048796270 hasConceptScore W3048796270C162324750 @default.
• W3048796270 hasConceptScore W3048796270C187653413 @default.
• W3048796270 hasConceptScore W3048796270C19499675 @default.
• W3048796270 hasConceptScore W3048796270C2775937380 @default.

• next