Matches in SemOpenAlex for { <https://semopenalex.org/work/W4385764785> ?p ?o ?g. }
Showing items 1 to 80 of
80
with 100 items per page.
- W4385764785 abstract "There are increasing demands for understanding deep neural networks' (DNNs) behavior spurred by growing security and/or transparency concerns. Due to multi-layer nonlinearity of the deep neural network architectures, explaining DNN predictions still remains as an open problem, preventing us from gaining a deeper understanding of the mechanisms. To enhance the explainability of DNNs, we estimate the input feature's attributions to the prediction task using divergence and flux. Inspired by the divergence theorem in vector analysis, we develop a novel Negative Flux Aggregation (NeFLAG) formulation and an efficient approximation algorithm to estimate attribution map. Unlike the previous techniques, ours doesn't rely on fitting a surrogate model nor need any path integration of gradients. Both qualitative and quantitative experiments demonstrate a superior performance of NeFLAG in generating more faithful attribution maps than the competing methods. Our code is available at https://github.com/xinli0928/NeFLAG." @default.
- W4385764785 created "2023-08-12" @default.
- W4385764785 creator A5009256505 @default.
- W4385764785 creator A5014432190 @default.
- W4385764785 creator A5022526821 @default.
- W4385764785 creator A5029858912 @default.
- W4385764785 creator A5054809100 @default.
- W4385764785 date "2023-08-01" @default.
- W4385764785 modified "2023-10-18" @default.
- W4385764785 title "Negative Flux Aggregation to Estimate Feature Attributions" @default.
- W4385764785 doi "https://doi.org/10.24963/ijcai.2023/50" @default.
- W4385764785 hasPublicationYear "2023" @default.
- W4385764785 type Work @default.
- W4385764785 citedByCount "0" @default.
- W4385764785 crossrefType "proceedings-article" @default.
- W4385764785 hasAuthorship W4385764785A5009256505 @default.
- W4385764785 hasAuthorship W4385764785A5014432190 @default.
- W4385764785 hasAuthorship W4385764785A5022526821 @default.
- W4385764785 hasAuthorship W4385764785A5029858912 @default.
- W4385764785 hasAuthorship W4385764785A5054809100 @default.
- W4385764785 hasBestOaLocation W43857647851 @default.
- W4385764785 hasConcept C11413529 @default.
- W4385764785 hasConcept C119857082 @default.
- W4385764785 hasConcept C124101348 @default.
- W4385764785 hasConcept C138885662 @default.
- W4385764785 hasConcept C143299363 @default.
- W4385764785 hasConcept C154945302 @default.
- W4385764785 hasConcept C15744967 @default.
- W4385764785 hasConcept C162324750 @default.
- W4385764785 hasConcept C177264268 @default.
- W4385764785 hasConcept C187736073 @default.
- W4385764785 hasConcept C199360897 @default.
- W4385764785 hasConcept C207390915 @default.
- W4385764785 hasConcept C2776401178 @default.
- W4385764785 hasConcept C2776760102 @default.
- W4385764785 hasConcept C2780233690 @default.
- W4385764785 hasConcept C2780451532 @default.
- W4385764785 hasConcept C38652104 @default.
- W4385764785 hasConcept C41008148 @default.
- W4385764785 hasConcept C41895202 @default.
- W4385764785 hasConcept C50644808 @default.
- W4385764785 hasConcept C77805123 @default.
- W4385764785 hasConceptScore W4385764785C11413529 @default.
- W4385764785 hasConceptScore W4385764785C119857082 @default.
- W4385764785 hasConceptScore W4385764785C124101348 @default.
- W4385764785 hasConceptScore W4385764785C138885662 @default.
- W4385764785 hasConceptScore W4385764785C143299363 @default.
- W4385764785 hasConceptScore W4385764785C154945302 @default.
- W4385764785 hasConceptScore W4385764785C15744967 @default.
- W4385764785 hasConceptScore W4385764785C162324750 @default.
- W4385764785 hasConceptScore W4385764785C177264268 @default.
- W4385764785 hasConceptScore W4385764785C187736073 @default.
- W4385764785 hasConceptScore W4385764785C199360897 @default.
- W4385764785 hasConceptScore W4385764785C207390915 @default.
- W4385764785 hasConceptScore W4385764785C2776401178 @default.
- W4385764785 hasConceptScore W4385764785C2776760102 @default.
- W4385764785 hasConceptScore W4385764785C2780233690 @default.
- W4385764785 hasConceptScore W4385764785C2780451532 @default.
- W4385764785 hasConceptScore W4385764785C38652104 @default.
- W4385764785 hasConceptScore W4385764785C41008148 @default.
- W4385764785 hasConceptScore W4385764785C41895202 @default.
- W4385764785 hasConceptScore W4385764785C50644808 @default.
- W4385764785 hasConceptScore W4385764785C77805123 @default.
- W4385764785 hasLocation W43857647851 @default.
- W4385764785 hasLocation W43857647852 @default.
- W4385764785 hasOpenAccess W4385764785 @default.
- W4385764785 hasPrimaryLocation W43857647851 @default.
- W4385764785 hasRelatedWork W1976996175 @default.
- W4385764785 hasRelatedWork W2081647779 @default.
- W4385764785 hasRelatedWork W2961085424 @default.
- W4385764785 hasRelatedWork W3046775127 @default.
- W4385764785 hasRelatedWork W3170094116 @default.
- W4385764785 hasRelatedWork W4285260836 @default.
- W4385764785 hasRelatedWork W4286629047 @default.
- W4385764785 hasRelatedWork W4306321456 @default.
- W4385764785 hasRelatedWork W4306674287 @default.
- W4385764785 hasRelatedWork W4224009465 @default.
- W4385764785 isParatext "false" @default.
- W4385764785 isRetracted "false" @default.
- W4385764785 workType "article" @default.