Matches in SemOpenAlex for { <https://semopenalex.org/work/W3131105417> ?p ?o ?g. }
- W3131105417 abstract "A neural network model of a differential equation, namely neural ODE, has enabled us to learn continuous-time dynamical systems and probabilistic distributions with a high accuracy. It uses the same network repeatedly during a numerical integration. Hence, the backpropagation algorithm requires a memory footprint proportional to the number of uses times the network size. This is true even if a checkpointing scheme divides the computational graph into sub-graphs. Otherwise, the adjoint method obtains a gradient by a numerical integration backward in time with a minimal memory footprint; however, it suffers from numerical errors. This study proposes the symplectic adjoint method, which obtains the exact gradient (up to rounding error) with a footprint proportional to the number of uses plus the network size. The experimental results demonstrate the symplectic adjoint method occupies the smallest footprint in most cases, functions faster in some cases, and is robust to a rounding error among competitive methods." @default.
- W3131105417 created "2021-03-01" @default.
- W3131105417 creator A5004524501 @default.
- W3131105417 creator A5007645171 @default.
- W3131105417 creator A5068667478 @default.
- W3131105417 date "2021-02-19" @default.
- W3131105417 modified "2023-09-28" @default.
- W3131105417 title "Symplectic Adjoint Method for Exact Gradient of Neural ODE with Minimal Memory" @default.
- W3131105417 cites W1498436455 @default.
- W3131105417 cites W1544770838 @default.
- W3131105417 cites W1815076433 @default.
- W3131105417 cites W2003379874 @default.
- W3131105417 cites W2037889116 @default.
- W3131105417 cites W2064675550 @default.
- W3131105417 cites W2082193626 @default.
- W3131105417 cites W2112796928 @default.
- W3131105417 cites W2137254413 @default.
- W3131105417 cites W2166751342 @default.
- W3131105417 cites W2194775991 @default.
- W3131105417 cites W2896457183 @default.
- W3131105417 cites W2899771611 @default.
- W3131105417 cites W2962695743 @default.
- W3131105417 cites W2963047245 @default.
- W3131105417 cites W2963090522 @default.
- W3131105417 cites W2963302407 @default.
- W3131105417 cites W2963684275 @default.
- W3131105417 cites W2963755523 @default.
- W3131105417 cites W2964020555 @default.
- W3131105417 cites W2964121744 @default.
- W3131105417 cites W2965543245 @default.
- W3131105417 cites W2986615800 @default.
- W3131105417 cites W2995774847 @default.
- W3131105417 cites W2999026783 @default.
- W3131105417 cites W3033752673 @default.
- W3131105417 cites W3099398368 @default.
- W3131105417 cites W3101274772 @default.
- W3131105417 cites W3107521724 @default.
- W3131105417 cites W3110996603 @default.
- W3131105417 cites W3126941302 @default.
- W3131105417 cites W3138232237 @default.
- W3131105417 cites W402518760 @default.
- W3131105417 cites W2962776038 @default.
- W3131105417 cites W3034652062 @default.
- W3131105417 cites W3193996508 @default.
- W3131105417 hasPublicationYear "2021" @default.
- W3131105417 type Work @default.
- W3131105417 sameAs 3131105417 @default.
- W3131105417 citedByCount "1" @default.
- W3131105417 countsByYear W31311054172020 @default.
- W3131105417 crossrefType "posted-content" @default.
- W3131105417 hasAuthorship W3131105417A5004524501 @default.
- W3131105417 hasAuthorship W3131105417A5007645171 @default.
- W3131105417 hasAuthorship W3131105417A5068667478 @default.
- W3131105417 hasConcept C111919701 @default.
- W3131105417 hasConcept C11413529 @default.
- W3131105417 hasConcept C127349201 @default.
- W3131105417 hasConcept C134306372 @default.
- W3131105417 hasConcept C136625980 @default.
- W3131105417 hasConcept C154945302 @default.
- W3131105417 hasConcept C155032097 @default.
- W3131105417 hasConcept C165551260 @default.
- W3131105417 hasConcept C168619227 @default.
- W3131105417 hasConcept C207646825 @default.
- W3131105417 hasConcept C28826006 @default.
- W3131105417 hasConcept C33923547 @default.
- W3131105417 hasConcept C34862557 @default.
- W3131105417 hasConcept C40265840 @default.
- W3131105417 hasConcept C41008148 @default.
- W3131105417 hasConcept C50644808 @default.
- W3131105417 hasConcept C61005703 @default.
- W3131105417 hasConcept C74912251 @default.
- W3131105417 hasConcept C78045399 @default.
- W3131105417 hasConceptScore W3131105417C111919701 @default.
- W3131105417 hasConceptScore W3131105417C11413529 @default.
- W3131105417 hasConceptScore W3131105417C127349201 @default.
- W3131105417 hasConceptScore W3131105417C134306372 @default.
- W3131105417 hasConceptScore W3131105417C136625980 @default.
- W3131105417 hasConceptScore W3131105417C154945302 @default.
- W3131105417 hasConceptScore W3131105417C155032097 @default.
- W3131105417 hasConceptScore W3131105417C165551260 @default.
- W3131105417 hasConceptScore W3131105417C168619227 @default.
- W3131105417 hasConceptScore W3131105417C207646825 @default.
- W3131105417 hasConceptScore W3131105417C28826006 @default.
- W3131105417 hasConceptScore W3131105417C33923547 @default.
- W3131105417 hasConceptScore W3131105417C34862557 @default.
- W3131105417 hasConceptScore W3131105417C40265840 @default.
- W3131105417 hasConceptScore W3131105417C41008148 @default.
- W3131105417 hasConceptScore W3131105417C50644808 @default.
- W3131105417 hasConceptScore W3131105417C61005703 @default.
- W3131105417 hasConceptScore W3131105417C74912251 @default.
- W3131105417 hasConceptScore W3131105417C78045399 @default.
- W3131105417 hasOpenAccess W3131105417 @default.
- W3131105417 hasRelatedWork W1503760862 @default.
- W3131105417 hasRelatedWork W1805759115 @default.
- W3131105417 hasRelatedWork W1824882229 @default.
- W3131105417 hasRelatedWork W1997497924 @default.
- W3131105417 hasRelatedWork W2024397048 @default.
- W3131105417 hasRelatedWork W2060306687 @default.
- W3131105417 hasRelatedWork W2081561302 @default.
- W3131105417 hasRelatedWork W2334661594 @default.