Matches in SemOpenAlex for { <https://semopenalex.org/work/W4386996974> ?p ?o ?g. }
- W4386996974 endingPage "453" @default.
- W4386996974 startingPage "453" @default.
- W4386996974 abstract "The Gaussian-radial-basis function neural network (GRBFNN) has been a popular choice for interpolation and classification. However, it is computationally intensive when the dimension of the input vector is high. To address this issue, we propose a new feedforward network-separable Gaussian neural network (SGNN) by taking advantage of the separable property of Gaussian-radial-basis functions, which splits input data into multiple columns and sequentially feeds them into parallel layers formed by uni-variate Gaussian functions. This structure reduces the number of neurons from O(Nd) of GRBFNN to O(dN), which exponentially improves the computational speed of SGNN and makes it scale linearly as the input dimension increases. In addition, SGNN can preserve the dominant subspace of the Hessian matrix of GRBFNN in gradient descent training, leading to a similar level of accuracy to GRBFNN. It is experimentally demonstrated that SGNN can achieve an acceleration of 100 times with a similar level of accuracy over GRBFNN on tri-variate function approximations. The SGNN also has better trainability and is more tuning-friendly than DNNs with RuLU and Sigmoid functions. For approximating functions with a complex geometry, SGNN can lead to results that are three orders of magnitude more accurate than those of a RuLU-DNN with twice the number of layers and the number of neurons per layer." @default.
- W4386996974 created "2023-09-25" @default.
- W4386996974 creator A5003255279 @default.
- W4386996974 creator A5017350937 @default.
- W4386996974 date "2023-09-22" @default.
- W4386996974 modified "2023-09-30" @default.
- W4386996974 title "Separable Gaussian Neural Networks: Structure, Analysis, and Function Approximations" @default.
- W4386996974 cites W1594640477 @default.
- W4386996974 cites W1999433604 @default.
- W4386996974 cites W2034374737 @default.
- W4386996974 cites W2073713860 @default.
- W4386996974 cites W2093022009 @default.
- W4386996974 cites W2113442785 @default.
- W4386996974 cites W2114865067 @default.
- W4386996974 cites W2116119284 @default.
- W4386996974 cites W2135894089 @default.
- W4386996974 cites W2137983211 @default.
- W4386996974 cites W2143956139 @default.
- W4386996974 cites W2155399784 @default.
- W4386996974 cites W2158426375 @default.
- W4386996974 cites W2171277043 @default.
- W4386996974 cites W2497532187 @default.
- W4386996974 cites W2582881728 @default.
- W4386996974 cites W2917723212 @default.
- W4386996974 cites W2968460293 @default.
- W4386996974 cites W3035024835 @default.
- W4386996974 cites W3125537303 @default.
- W4386996974 cites W4211093736 @default.
- W4386996974 doi "https://doi.org/10.3390/a16100453" @default.
- W4386996974 hasPublicationYear "2023" @default.
- W4386996974 type Work @default.
- W4386996974 citedByCount "0" @default.
- W4386996974 crossrefType "journal-article" @default.
- W4386996974 hasAuthorship W4386996974A5003255279 @default.
- W4386996974 hasAuthorship W4386996974A5017350937 @default.
- W4386996974 hasBestOaLocation W43869969741 @default.
- W4386996974 hasConcept C104114177 @default.
- W4386996974 hasConcept C11413529 @default.
- W4386996974 hasConcept C114614502 @default.
- W4386996974 hasConcept C121332964 @default.
- W4386996974 hasConcept C134306372 @default.
- W4386996974 hasConcept C137800194 @default.
- W4386996974 hasConcept C14036430 @default.
- W4386996974 hasConcept C153258448 @default.
- W4386996974 hasConcept C154945302 @default.
- W4386996974 hasConcept C163716315 @default.
- W4386996974 hasConcept C203616005 @default.
- W4386996974 hasConcept C28826006 @default.
- W4386996974 hasConcept C33676613 @default.
- W4386996974 hasConcept C33923547 @default.
- W4386996974 hasConcept C41008148 @default.
- W4386996974 hasConcept C50644808 @default.
- W4386996974 hasConcept C5917680 @default.
- W4386996974 hasConcept C62520636 @default.
- W4386996974 hasConcept C70710897 @default.
- W4386996974 hasConcept C7218915 @default.
- W4386996974 hasConcept C78458016 @default.
- W4386996974 hasConcept C81388566 @default.
- W4386996974 hasConcept C86803240 @default.
- W4386996974 hasConcept C98856871 @default.
- W4386996974 hasConceptScore W4386996974C104114177 @default.
- W4386996974 hasConceptScore W4386996974C11413529 @default.
- W4386996974 hasConceptScore W4386996974C114614502 @default.
- W4386996974 hasConceptScore W4386996974C121332964 @default.
- W4386996974 hasConceptScore W4386996974C134306372 @default.
- W4386996974 hasConceptScore W4386996974C137800194 @default.
- W4386996974 hasConceptScore W4386996974C14036430 @default.
- W4386996974 hasConceptScore W4386996974C153258448 @default.
- W4386996974 hasConceptScore W4386996974C154945302 @default.
- W4386996974 hasConceptScore W4386996974C163716315 @default.
- W4386996974 hasConceptScore W4386996974C203616005 @default.
- W4386996974 hasConceptScore W4386996974C28826006 @default.
- W4386996974 hasConceptScore W4386996974C33676613 @default.
- W4386996974 hasConceptScore W4386996974C33923547 @default.
- W4386996974 hasConceptScore W4386996974C41008148 @default.
- W4386996974 hasConceptScore W4386996974C50644808 @default.
- W4386996974 hasConceptScore W4386996974C5917680 @default.
- W4386996974 hasConceptScore W4386996974C62520636 @default.
- W4386996974 hasConceptScore W4386996974C70710897 @default.
- W4386996974 hasConceptScore W4386996974C7218915 @default.
- W4386996974 hasConceptScore W4386996974C78458016 @default.
- W4386996974 hasConceptScore W4386996974C81388566 @default.
- W4386996974 hasConceptScore W4386996974C86803240 @default.
- W4386996974 hasConceptScore W4386996974C98856871 @default.
- W4386996974 hasIssue "10" @default.
- W4386996974 hasLocation W43869969741 @default.
- W4386996974 hasOpenAccess W4386996974 @default.
- W4386996974 hasPrimaryLocation W43869969741 @default.
- W4386996974 hasRelatedWork W1844997853 @default.
- W4386996974 hasRelatedWork W1992087623 @default.
- W4386996974 hasRelatedWork W2057269233 @default.
- W4386996974 hasRelatedWork W2103176524 @default.
- W4386996974 hasRelatedWork W2356056792 @default.
- W4386996974 hasRelatedWork W256613907 @default.
- W4386996974 hasRelatedWork W3006617118 @default.
- W4386996974 hasRelatedWork W3149446626 @default.
- W4386996974 hasRelatedWork W4281553196 @default.
- W4386996974 hasRelatedWork W2622138135 @default.