Matches in SemOpenAlex for { <https://semopenalex.org/work/W4300717047> ?p ?o ?g. }
Showing items 1 to 78 of
78
with 100 items per page.
- W4300717047 endingPage "103803" @default.
- W4300717047 startingPage "103803" @default.
- W4300717047 abstract "Kernel methods are powerful tools to capture nonlinear patterns behind given data but often lead to poor performance on complicated tasks compared to convolutional neural networks. The reason is that kernel methods are still shallow and fully connected models, failing to reveal hierarchical features and local interdependencies. In this paper, to acquire hierarchical and local knowledge, we incorporate kernel methods with deep architectures and convolutional operators in a spectral kernel learning framework. Based on the inverse Fourier transform and Rademacher complexity theory, we provide the generalization error bounds for the proposed model and prove that under suitable initialization, deeper networks lead to tighter error bounds. Inspired by theoretical findings, we finally completed the convolutional spectral kernel network (CSKN) with two additional regularizers and an initialization strategy. Extensive ablation results validate the effectiveness of non-stationary spectral kernel, multiple layers, additional regularizers, and the convolutional filters, which coincide with our theoretical findings. We further devise a VGG-type 8-layers CSKN, and it outperforms the existing kernel-based networks and popular CNN models on the medium-sized image classification tasks." @default.
- W4300717047 created "2022-10-04" @default.
- W4300717047 creator A5023783253 @default.
- W4300717047 creator A5028071674 @default.
- W4300717047 creator A5077114895 @default.
- W4300717047 date "2022-12-01" @default.
- W4300717047 modified "2023-09-23" @default.
- W4300717047 title "Convolutional spectral kernel learning with generalization guarantees" @default.
- W4300717047 cites W1994906459 @default.
- W4300717047 cites W2009784682 @default.
- W4300717047 cites W2087258353 @default.
- W4300717047 cites W2112796928 @default.
- W4300717047 cites W2962957157 @default.
- W4300717047 cites W2963626582 @default.
- W4300717047 cites W3000731189 @default.
- W4300717047 cites W3002335888 @default.
- W4300717047 cites W3100743579 @default.
- W4300717047 doi "https://doi.org/10.1016/j.artint.2022.103803" @default.
- W4300717047 hasPublicationYear "2022" @default.
- W4300717047 type Work @default.
- W4300717047 citedByCount "0" @default.
- W4300717047 crossrefType "journal-article" @default.
- W4300717047 hasAuthorship W4300717047A5023783253 @default.
- W4300717047 hasAuthorship W4300717047A5028071674 @default.
- W4300717047 hasAuthorship W4300717047A5077114895 @default.
- W4300717047 hasConcept C11413529 @default.
- W4300717047 hasConcept C114466953 @default.
- W4300717047 hasConcept C114614502 @default.
- W4300717047 hasConcept C119857082 @default.
- W4300717047 hasConcept C122280245 @default.
- W4300717047 hasConcept C12267149 @default.
- W4300717047 hasConcept C134306372 @default.
- W4300717047 hasConcept C134517425 @default.
- W4300717047 hasConcept C153180895 @default.
- W4300717047 hasConcept C154945302 @default.
- W4300717047 hasConcept C177148314 @default.
- W4300717047 hasConcept C199360897 @default.
- W4300717047 hasConcept C2776879701 @default.
- W4300717047 hasConcept C33923547 @default.
- W4300717047 hasConcept C41008148 @default.
- W4300717047 hasConcept C74193536 @default.
- W4300717047 hasConcept C81363708 @default.
- W4300717047 hasConceptScore W4300717047C11413529 @default.
- W4300717047 hasConceptScore W4300717047C114466953 @default.
- W4300717047 hasConceptScore W4300717047C114614502 @default.
- W4300717047 hasConceptScore W4300717047C119857082 @default.
- W4300717047 hasConceptScore W4300717047C122280245 @default.
- W4300717047 hasConceptScore W4300717047C12267149 @default.
- W4300717047 hasConceptScore W4300717047C134306372 @default.
- W4300717047 hasConceptScore W4300717047C134517425 @default.
- W4300717047 hasConceptScore W4300717047C153180895 @default.
- W4300717047 hasConceptScore W4300717047C154945302 @default.
- W4300717047 hasConceptScore W4300717047C177148314 @default.
- W4300717047 hasConceptScore W4300717047C199360897 @default.
- W4300717047 hasConceptScore W4300717047C2776879701 @default.
- W4300717047 hasConceptScore W4300717047C33923547 @default.
- W4300717047 hasConceptScore W4300717047C41008148 @default.
- W4300717047 hasConceptScore W4300717047C74193536 @default.
- W4300717047 hasConceptScore W4300717047C81363708 @default.
- W4300717047 hasLocation W43007170471 @default.
- W4300717047 hasOpenAccess W4300717047 @default.
- W4300717047 hasPrimaryLocation W43007170471 @default.
- W4300717047 hasRelatedWork W2043864454 @default.
- W4300717047 hasRelatedWork W2092483655 @default.
- W4300717047 hasRelatedWork W2188831877 @default.
- W4300717047 hasRelatedWork W2206558667 @default.
- W4300717047 hasRelatedWork W2604913466 @default.
- W4300717047 hasRelatedWork W2725311638 @default.
- W4300717047 hasRelatedWork W2828181497 @default.
- W4300717047 hasRelatedWork W2970160020 @default.
- W4300717047 hasRelatedWork W4291669689 @default.
- W4300717047 hasRelatedWork W4300717047 @default.
- W4300717047 hasVolume "313" @default.
- W4300717047 isParatext "false" @default.
- W4300717047 isRetracted "false" @default.
- W4300717047 workType "article" @default.