Matches in SemOpenAlex for { <https://semopenalex.org/work/W4287366385> ?p ?o ?g. }
Showing items 1 to 81 of
81
with 100 items per page.
- W4287366385 abstract "Approximation theorists have established best-in-class optimal approximation rates of deep neural networks by utilizing their ability to simultaneously emulate partitions of unity and monomials. Motivated by this, we propose partition of unity networks (POUnets) which incorporate these elements directly into the architecture. Classification architectures of the type used to learn probability measures are used to build a meshfree partition of space, while polynomial spaces with learnable coefficients are associated to each partition. The resulting hp-element-like approximation allows use of a fast least-squares optimizer, and the resulting architecture size need not scale exponentially with spatial dimension, breaking the curse of dimensionality. An abstract approximation result establishes desirable properties to guide network design. Numerical results for two choices of architecture demonstrate that POUnets yield hp-convergence for smooth functions and consistently outperform MLPs for piecewise polynomial functions with large numbers of discontinuities." @default.
- W4287366385 created "2022-07-25" @default.
- W4287366385 creator A5003125158 @default.
- W4287366385 creator A5012880865 @default.
- W4287366385 creator A5052883660 @default.
- W4287366385 creator A5056314900 @default.
- W4287366385 creator A5062691605 @default.
- W4287366385 date "2021-01-27" @default.
- W4287366385 modified "2023-09-26" @default.
- W4287366385 title "Partition of unity networks: deep hp-approximation" @default.
- W4287366385 doi "https://doi.org/10.48550/arxiv.2101.11256" @default.
- W4287366385 hasPublicationYear "2021" @default.
- W4287366385 type Work @default.
- W4287366385 citedByCount "1" @default.
- W4287366385 countsByYear W42873663852023 @default.
- W4287366385 crossrefType "posted-content" @default.
- W4287366385 hasAuthorship W4287366385A5003125158 @default.
- W4287366385 hasAuthorship W4287366385A5012880865 @default.
- W4287366385 hasAuthorship W4287366385A5052883660 @default.
- W4287366385 hasAuthorship W4287366385A5056314900 @default.
- W4287366385 hasAuthorship W4287366385A5062691605 @default.
- W4287366385 hasBestOaLocation W42873663851 @default.
- W4287366385 hasConcept C102205669 @default.
- W4287366385 hasConcept C105795698 @default.
- W4287366385 hasConcept C111030470 @default.
- W4287366385 hasConcept C11252640 @default.
- W4287366385 hasConcept C114614502 @default.
- W4287366385 hasConcept C118615104 @default.
- W4287366385 hasConcept C121332964 @default.
- W4287366385 hasConcept C126255220 @default.
- W4287366385 hasConcept C134306372 @default.
- W4287366385 hasConcept C135628077 @default.
- W4287366385 hasConcept C154945302 @default.
- W4287366385 hasConcept C164660894 @default.
- W4287366385 hasConcept C2780860992 @default.
- W4287366385 hasConcept C28826006 @default.
- W4287366385 hasConcept C33923547 @default.
- W4287366385 hasConcept C41008148 @default.
- W4287366385 hasConcept C42812 @default.
- W4287366385 hasConcept C50644808 @default.
- W4287366385 hasConcept C77088390 @default.
- W4287366385 hasConcept C90119067 @default.
- W4287366385 hasConcept C97355855 @default.
- W4287366385 hasConceptScore W4287366385C102205669 @default.
- W4287366385 hasConceptScore W4287366385C105795698 @default.
- W4287366385 hasConceptScore W4287366385C111030470 @default.
- W4287366385 hasConceptScore W4287366385C11252640 @default.
- W4287366385 hasConceptScore W4287366385C114614502 @default.
- W4287366385 hasConceptScore W4287366385C118615104 @default.
- W4287366385 hasConceptScore W4287366385C121332964 @default.
- W4287366385 hasConceptScore W4287366385C126255220 @default.
- W4287366385 hasConceptScore W4287366385C134306372 @default.
- W4287366385 hasConceptScore W4287366385C135628077 @default.
- W4287366385 hasConceptScore W4287366385C154945302 @default.
- W4287366385 hasConceptScore W4287366385C164660894 @default.
- W4287366385 hasConceptScore W4287366385C2780860992 @default.
- W4287366385 hasConceptScore W4287366385C28826006 @default.
- W4287366385 hasConceptScore W4287366385C33923547 @default.
- W4287366385 hasConceptScore W4287366385C41008148 @default.
- W4287366385 hasConceptScore W4287366385C42812 @default.
- W4287366385 hasConceptScore W4287366385C50644808 @default.
- W4287366385 hasConceptScore W4287366385C77088390 @default.
- W4287366385 hasConceptScore W4287366385C90119067 @default.
- W4287366385 hasConceptScore W4287366385C97355855 @default.
- W4287366385 hasLocation W42873663851 @default.
- W4287366385 hasLocation W42873663852 @default.
- W4287366385 hasOpenAccess W4287366385 @default.
- W4287366385 hasPrimaryLocation W42873663851 @default.
- W4287366385 hasRelatedWork W1968915506 @default.
- W4287366385 hasRelatedWork W2039581508 @default.
- W4287366385 hasRelatedWork W2057954107 @default.
- W4287366385 hasRelatedWork W2072084563 @default.
- W4287366385 hasRelatedWork W2074195844 @default.
- W4287366385 hasRelatedWork W2989875800 @default.
- W4287366385 hasRelatedWork W3123521181 @default.
- W4287366385 hasRelatedWork W3210910181 @default.
- W4287366385 hasRelatedWork W4248912794 @default.
- W4287366385 hasRelatedWork W4287366385 @default.
- W4287366385 isParatext "false" @default.
- W4287366385 isRetracted "false" @default.
- W4287366385 workType "article" @default.