Matches in SemOpenAlex for { <https://semopenalex.org/work/W4225928049> ?p ?o ?g. }
Showing items 1 to 66 of
66
with 100 items per page.
- W4225928049 abstract "Existing methods of pruning deep neural networks focus on removing unnecessary parameters of the trained network and fine tuning the model afterwards to find a good solution that recovers the initial performance of the trained model. Unlike other works, our method pays special attention to the quality of the solution in the compressed model and inference computation time by pruning neurons. The proposed algorithm directs the parameters of the compressed model toward a flatter solution by exploring the spectral radius of Hessian which results in better generalization on unseen data. Moreover, the method does not work with a pre-trained network and performs training and pruning simultaneously. Our result shows that it improves the state-of-the-art results on neuron compression. The method is able to achieve very small networks with small accuracy degradation across different neural network models." @default.
- W4225928049 created "2022-05-05" @default.
- W4225928049 creator A5013049879 @default.
- W4225928049 creator A5022228891 @default.
- W4225928049 date "2021-11-16" @default.
- W4225928049 modified "2023-09-29" @default.
- W4225928049 title "Neuron-based Pruning of Deep Neural Networks with Better Generalization using Kronecker Factored Curvature Approximation" @default.
- W4225928049 hasPublicationYear "2021" @default.
- W4225928049 type Work @default.
- W4225928049 citedByCount "0" @default.
- W4225928049 crossrefType "posted-content" @default.
- W4225928049 hasAuthorship W4225928049A5013049879 @default.
- W4225928049 hasAuthorship W4225928049A5022228891 @default.
- W4225928049 hasBestOaLocation W42259280491 @default.
- W4225928049 hasConcept C108010975 @default.
- W4225928049 hasConcept C11413529 @default.
- W4225928049 hasConcept C120665830 @default.
- W4225928049 hasConcept C121332964 @default.
- W4225928049 hasConcept C134306372 @default.
- W4225928049 hasConcept C154945302 @default.
- W4225928049 hasConcept C177148314 @default.
- W4225928049 hasConcept C192209626 @default.
- W4225928049 hasConcept C203616005 @default.
- W4225928049 hasConcept C2776214188 @default.
- W4225928049 hasConcept C28826006 @default.
- W4225928049 hasConcept C2984842247 @default.
- W4225928049 hasConcept C33923547 @default.
- W4225928049 hasConcept C41008148 @default.
- W4225928049 hasConcept C45374587 @default.
- W4225928049 hasConcept C50644808 @default.
- W4225928049 hasConcept C6557445 @default.
- W4225928049 hasConcept C86803240 @default.
- W4225928049 hasConceptScore W4225928049C108010975 @default.
- W4225928049 hasConceptScore W4225928049C11413529 @default.
- W4225928049 hasConceptScore W4225928049C120665830 @default.
- W4225928049 hasConceptScore W4225928049C121332964 @default.
- W4225928049 hasConceptScore W4225928049C134306372 @default.
- W4225928049 hasConceptScore W4225928049C154945302 @default.
- W4225928049 hasConceptScore W4225928049C177148314 @default.
- W4225928049 hasConceptScore W4225928049C192209626 @default.
- W4225928049 hasConceptScore W4225928049C203616005 @default.
- W4225928049 hasConceptScore W4225928049C2776214188 @default.
- W4225928049 hasConceptScore W4225928049C28826006 @default.
- W4225928049 hasConceptScore W4225928049C2984842247 @default.
- W4225928049 hasConceptScore W4225928049C33923547 @default.
- W4225928049 hasConceptScore W4225928049C41008148 @default.
- W4225928049 hasConceptScore W4225928049C45374587 @default.
- W4225928049 hasConceptScore W4225928049C50644808 @default.
- W4225928049 hasConceptScore W4225928049C6557445 @default.
- W4225928049 hasConceptScore W4225928049C86803240 @default.
- W4225928049 hasLocation W42259280491 @default.
- W4225928049 hasOpenAccess W4225928049 @default.
- W4225928049 hasPrimaryLocation W42259280491 @default.
- W4225928049 hasRelatedWork W13531899 @default.
- W4225928049 hasRelatedWork W1424933 @default.
- W4225928049 hasRelatedWork W15476695 @default.
- W4225928049 hasRelatedWork W15833439 @default.
- W4225928049 hasRelatedWork W2180734 @default.
- W4225928049 hasRelatedWork W311676 @default.
- W4225928049 hasRelatedWork W3803293 @default.
- W4225928049 hasRelatedWork W6661183 @default.
- W4225928049 hasRelatedWork W67111 @default.
- W4225928049 hasRelatedWork W7497093 @default.
- W4225928049 isParatext "false" @default.
- W4225928049 isRetracted "false" @default.
- W4225928049 workType "article" @default.