Matches in SemOpenAlex for { <https://semopenalex.org/work/W3199976834> ?p ?o ?g. }
Showing items 1 to 97 of
97
with 100 items per page.
- W3199976834 abstract "We propose a convenient matrix-free neural architecture for the multigrid method. The architecture is simple enough to be implemented in less than fifty lines of code, yet it encompasses a large number of distinct multigrid solvers. We argue that a fixed neural network without dense layers can not realize an efficient iterative method. Because of that, standard training protocols do not lead to competitive solvers. To overcome this difficulty, we use parameter sharing and serialization of layers. The resulting network can be trained on linear problems with thousands of unknowns and retains its efficiency on problems with millions of unknowns. From the point of view of numerical linear algebra network's training corresponds to finding optimal smoothers for the geometric multigrid method. We demonstrate our approach on a few second-order elliptic equations. For tested linear systems, we obtain from two to five times smaller spectral radius of the error propagation matrix compare to a basic linear multigrid with Jacobi smoother." @default.
- W3199976834 created "2021-09-27" @default.
- W3199976834 creator A5013778722 @default.
- W3199976834 date "2021-07-18" @default.
- W3199976834 modified "2023-10-16" @default.
- W3199976834 title "Neural Multigrid Architectures" @default.
- W3199976834 cites W1506342804 @default.
- W3199976834 cites W1990671420 @default.
- W3199976834 cites W2018525258 @default.
- W3199976834 cites W2036779012 @default.
- W3199976834 cites W2053486661 @default.
- W3199976834 cites W2057450058 @default.
- W3199976834 cites W2139923370 @default.
- W3199976834 cites W2192203593 @default.
- W3199976834 cites W2912759508 @default.
- W3199976834 cites W2980085231 @default.
- W3199976834 cites W3040239497 @default.
- W3199976834 cites W4211094263 @default.
- W3199976834 cites W4292025407 @default.
- W3199976834 cites W4292069163 @default.
- W3199976834 doi "https://doi.org/10.1109/ijcnn52387.2021.9533736" @default.
- W3199976834 hasPublicationYear "2021" @default.
- W3199976834 type Work @default.
- W3199976834 sameAs 3199976834 @default.
- W3199976834 citedByCount "2" @default.
- W3199976834 countsByYear W31999768342022 @default.
- W3199976834 crossrefType "proceedings-article" @default.
- W3199976834 hasAuthorship W3199976834A5013778722 @default.
- W3199976834 hasConcept C106487976 @default.
- W3199976834 hasConcept C111919701 @default.
- W3199976834 hasConcept C11413529 @default.
- W3199976834 hasConcept C121332964 @default.
- W3199976834 hasConcept C126255220 @default.
- W3199976834 hasConcept C134306372 @default.
- W3199976834 hasConcept C137119250 @default.
- W3199976834 hasConcept C139352143 @default.
- W3199976834 hasConcept C140532419 @default.
- W3199976834 hasConcept C154945302 @default.
- W3199976834 hasConcept C155512373 @default.
- W3199976834 hasConcept C158693339 @default.
- W3199976834 hasConcept C159985019 @default.
- W3199976834 hasConcept C163834973 @default.
- W3199976834 hasConcept C192562407 @default.
- W3199976834 hasConcept C2524010 @default.
- W3199976834 hasConcept C28826006 @default.
- W3199976834 hasConcept C33923547 @default.
- W3199976834 hasConcept C41008148 @default.
- W3199976834 hasConcept C50644808 @default.
- W3199976834 hasConcept C52723943 @default.
- W3199976834 hasConcept C60866291 @default.
- W3199976834 hasConcept C62520636 @default.
- W3199976834 hasConcept C6802819 @default.
- W3199976834 hasConcept C78045399 @default.
- W3199976834 hasConcept C94523830 @default.
- W3199976834 hasConceptScore W3199976834C106487976 @default.
- W3199976834 hasConceptScore W3199976834C111919701 @default.
- W3199976834 hasConceptScore W3199976834C11413529 @default.
- W3199976834 hasConceptScore W3199976834C121332964 @default.
- W3199976834 hasConceptScore W3199976834C126255220 @default.
- W3199976834 hasConceptScore W3199976834C134306372 @default.
- W3199976834 hasConceptScore W3199976834C137119250 @default.
- W3199976834 hasConceptScore W3199976834C139352143 @default.
- W3199976834 hasConceptScore W3199976834C140532419 @default.
- W3199976834 hasConceptScore W3199976834C154945302 @default.
- W3199976834 hasConceptScore W3199976834C155512373 @default.
- W3199976834 hasConceptScore W3199976834C158693339 @default.
- W3199976834 hasConceptScore W3199976834C159985019 @default.
- W3199976834 hasConceptScore W3199976834C163834973 @default.
- W3199976834 hasConceptScore W3199976834C192562407 @default.
- W3199976834 hasConceptScore W3199976834C2524010 @default.
- W3199976834 hasConceptScore W3199976834C28826006 @default.
- W3199976834 hasConceptScore W3199976834C33923547 @default.
- W3199976834 hasConceptScore W3199976834C41008148 @default.
- W3199976834 hasConceptScore W3199976834C50644808 @default.
- W3199976834 hasConceptScore W3199976834C52723943 @default.
- W3199976834 hasConceptScore W3199976834C60866291 @default.
- W3199976834 hasConceptScore W3199976834C62520636 @default.
- W3199976834 hasConceptScore W3199976834C6802819 @default.
- W3199976834 hasConceptScore W3199976834C78045399 @default.
- W3199976834 hasConceptScore W3199976834C94523830 @default.
- W3199976834 hasLocation W31999768341 @default.
- W3199976834 hasOpenAccess W3199976834 @default.
- W3199976834 hasPrimaryLocation W31999768341 @default.
- W3199976834 hasRelatedWork W1989528974 @default.
- W3199976834 hasRelatedWork W2026655828 @default.
- W3199976834 hasRelatedWork W2053827494 @default.
- W3199976834 hasRelatedWork W2129414784 @default.
- W3199976834 hasRelatedWork W2162572068 @default.
- W3199976834 hasRelatedWork W2315186575 @default.
- W3199976834 hasRelatedWork W2482977397 @default.
- W3199976834 hasRelatedWork W3023718732 @default.
- W3199976834 hasRelatedWork W3199976834 @default.
- W3199976834 hasRelatedWork W422177830 @default.
- W3199976834 isParatext "false" @default.
- W3199976834 isRetracted "false" @default.
- W3199976834 magId "3199976834" @default.
- W3199976834 workType "article" @default.