Matches in SemOpenAlex for { <https://semopenalex.org/work/W2949271805> ?p ?o ?g. }
Showing items 1 to 97 of
97
with 100 items per page.
- W2949271805 abstract "It has been observed that residual networks can be viewed as the explicit Euler discretization of an Ordinary Differential Equation (ODE). This observation motivated the introduction of so-called Neural ODEs, which allow more general discretization schemes with adaptive time stepping. Here, we propose ANODEV2, which is an extension of this approach that also allows evolution of the neural network parameters, in a coupled ODE-based formulation. The Neural ODE method introduced earlier is in fact a special case of this new more general framework. We present the formulation of ANODEV2, derive optimality conditions, and implement a coupled reaction-diffusion-advection version of this framework in PyTorch. We present empirical results using several different configurations of ANODEV2, testing them on multiple models on CIFAR-10. We report results showing that this coupled ODE-based framework is indeed trainable, and that it achieves higher accuracy, as compared to the baseline models as well as the recently-proposed Neural ODE approach." @default.
- W2949271805 created "2019-06-27" @default.
- W2949271805 creator A5025376043 @default.
- W2949271805 creator A5033006662 @default.
- W2949271805 creator A5033384149 @default.
- W2949271805 creator A5044137409 @default.
- W2949271805 creator A5047285420 @default.
- W2949271805 creator A5070987734 @default.
- W2949271805 creator A5072427753 @default.
- W2949271805 date "2019-06-09" @default.
- W2949271805 modified "2023-10-16" @default.
- W2949271805 title "ANODEV2: A Coupled Neural ODE Evolution Framework" @default.
- W2949271805 cites W1891300059 @default.
- W2949271805 cites W1985072454 @default.
- W2949271805 cites W2010769244 @default.
- W2949271805 cites W2020399841 @default.
- W2949271805 cites W2081547991 @default.
- W2949271805 cites W2089217417 @default.
- W2949271805 cites W2119814172 @default.
- W2949271805 cites W2138321920 @default.
- W2949271805 cites W2194775991 @default.
- W2949271805 cites W2302255633 @default.
- W2949271805 cites W2600297185 @default.
- W2949271805 cites W26461328 @default.
- W2949271805 cites W2903186160 @default.
- W2949271805 cites W2930930439 @default.
- W2949271805 cites W2963004026 @default.
- W2949271805 cites W2963302407 @default.
- W2949271805 cites W2963588848 @default.
- W2949271805 cites W2963755523 @default.
- W2949271805 cites W2965543245 @default.
- W2949271805 cites W2974916071 @default.
- W2949271805 cites W298987871 @default.
- W2949271805 cites W3098011980 @default.
- W2949271805 cites W91221188 @default.
- W2949271805 cites W996746306 @default.
- W2949271805 doi "https://doi.org/10.48550/arxiv.1906.04596" @default.
- W2949271805 hasPublicationYear "2019" @default.
- W2949271805 type Work @default.
- W2949271805 sameAs 2949271805 @default.
- W2949271805 citedByCount "18" @default.
- W2949271805 countsByYear W29492718052019 @default.
- W2949271805 countsByYear W29492718052020 @default.
- W2949271805 countsByYear W29492718052021 @default.
- W2949271805 countsByYear W29492718052022 @default.
- W2949271805 crossrefType "posted-content" @default.
- W2949271805 hasAuthorship W2949271805A5025376043 @default.
- W2949271805 hasAuthorship W2949271805A5033006662 @default.
- W2949271805 hasAuthorship W2949271805A5033384149 @default.
- W2949271805 hasAuthorship W2949271805A5044137409 @default.
- W2949271805 hasAuthorship W2949271805A5047285420 @default.
- W2949271805 hasAuthorship W2949271805A5070987734 @default.
- W2949271805 hasAuthorship W2949271805A5072427753 @default.
- W2949271805 hasBestOaLocation W29492718051 @default.
- W2949271805 hasConcept C11413529 @default.
- W2949271805 hasConcept C126255220 @default.
- W2949271805 hasConcept C134306372 @default.
- W2949271805 hasConcept C154945302 @default.
- W2949271805 hasConcept C155512373 @default.
- W2949271805 hasConcept C28826006 @default.
- W2949271805 hasConcept C33923547 @default.
- W2949271805 hasConcept C34862557 @default.
- W2949271805 hasConcept C41008148 @default.
- W2949271805 hasConcept C50644808 @default.
- W2949271805 hasConcept C51544822 @default.
- W2949271805 hasConcept C73000952 @default.
- W2949271805 hasConcept C78045399 @default.
- W2949271805 hasConceptScore W2949271805C11413529 @default.
- W2949271805 hasConceptScore W2949271805C126255220 @default.
- W2949271805 hasConceptScore W2949271805C134306372 @default.
- W2949271805 hasConceptScore W2949271805C154945302 @default.
- W2949271805 hasConceptScore W2949271805C155512373 @default.
- W2949271805 hasConceptScore W2949271805C28826006 @default.
- W2949271805 hasConceptScore W2949271805C33923547 @default.
- W2949271805 hasConceptScore W2949271805C34862557 @default.
- W2949271805 hasConceptScore W2949271805C41008148 @default.
- W2949271805 hasConceptScore W2949271805C50644808 @default.
- W2949271805 hasConceptScore W2949271805C51544822 @default.
- W2949271805 hasConceptScore W2949271805C73000952 @default.
- W2949271805 hasConceptScore W2949271805C78045399 @default.
- W2949271805 hasLocation W29492718051 @default.
- W2949271805 hasOpenAccess W2949271805 @default.
- W2949271805 hasPrimaryLocation W29492718051 @default.
- W2949271805 hasRelatedWork W1553476531 @default.
- W2949271805 hasRelatedWork W2136804894 @default.
- W2949271805 hasRelatedWork W2239576720 @default.
- W2949271805 hasRelatedWork W2769297273 @default.
- W2949271805 hasRelatedWork W2974753373 @default.
- W2949271805 hasRelatedWork W2979394879 @default.
- W2949271805 hasRelatedWork W3120553144 @default.
- W2949271805 hasRelatedWork W3135923918 @default.
- W2949271805 hasRelatedWork W3210092022 @default.
- W2949271805 hasRelatedWork W4282978001 @default.
- W2949271805 isParatext "false" @default.
- W2949271805 isRetracted "false" @default.
- W2949271805 magId "2949271805" @default.
- W2949271805 workType "article" @default.