Matches in SemOpenAlex for { <https://semopenalex.org/work/W3018301625> ?p ?o ?g. }
- W3018301625 endingPage "103682" @default.
- W3018301625 startingPage "103682" @default.
- W3018301625 abstract "We propose a discrimination-aware learning method to improve both the accuracy and fairness of biased face recognition algorithms. The most popular face recognition benchmarks assume a distribution of subjects without paying much attention to their demographic attributes. In this work, we perform a comprehensive discrimination-aware experimentation of deep learning-based face recognition. We also propose a notational framework for algorithmic discrimination with application to face biometrics. The experiments include three popular face recognition models and three public databases composed of 64,000 identities from different demographic groups characterized by sex and ethnicity. We experimentally show that learning processes based on the most used face databases have led to popular pre-trained deep face models that present evidence of strong algorithmic discrimination. Finally, we propose a discrimination-aware learning method, Sensitive Loss, based on the popular triplet loss function and a sensitive triplet generator. Our approach works as an add-on to pre-trained networks and is used to improve their performance in terms of average accuracy and fairness. The method shows results comparable to state-of-the-art de-biasing networks and represents a step forward to prevent discriminatory automatic systems." @default.
- W3018301625 created "2020-05-01" @default.
- W3018301625 creator A5004265472 @default.
- W3018301625 creator A5009232244 @default.
- W3018301625 creator A5017095669 @default.
- W3018301625 creator A5068081385 @default.
- W3018301625 creator A5083125799 @default.
- W3018301625 creator A5085139454 @default.
- W3018301625 date "2022-04-01" @default.
- W3018301625 modified "2023-09-30" @default.
- W3018301625 title "Sensitive loss: Improving accuracy and fairness of face representations with discrimination-aware deep learning" @default.
- W3018301625 cites W1978642336 @default.
- W3018301625 cites W2097246321 @default.
- W3018301625 cites W2126448884 @default.
- W3018301625 cites W2144942180 @default.
- W3018301625 cites W2741881058 @default.
- W3018301625 cites W2776637495 @default.
- W3018301625 cites W2784025535 @default.
- W3018301625 cites W2912376035 @default.
- W3018301625 cites W2947709300 @default.
- W3018301625 cites W2949386440 @default.
- W3018301625 cites W2950068534 @default.
- W3018301625 cites W2963212406 @default.
- W3018301625 cites W3009306238 @default.
- W3018301625 cites W3010093963 @default.
- W3018301625 cites W3035232573 @default.
- W3018301625 cites W3048511172 @default.
- W3018301625 cites W3101998545 @default.
- W3018301625 cites W3198878858 @default.
- W3018301625 cites W3199041638 @default.
- W3018301625 cites W3212169258 @default.
- W3018301625 cites W4211017739 @default.
- W3018301625 cites W4239774085 @default.
- W3018301625 doi "https://doi.org/10.1016/j.artint.2022.103682" @default.
- W3018301625 hasPublicationYear "2022" @default.
- W3018301625 type Work @default.
- W3018301625 sameAs 3018301625 @default.
- W3018301625 citedByCount "27" @default.
- W3018301625 countsByYear W30183016252020 @default.
- W3018301625 countsByYear W30183016252021 @default.
- W3018301625 countsByYear W30183016252022 @default.
- W3018301625 countsByYear W30183016252023 @default.
- W3018301625 crossrefType "journal-article" @default.
- W3018301625 hasAuthorship W3018301625A5004265472 @default.
- W3018301625 hasAuthorship W3018301625A5009232244 @default.
- W3018301625 hasAuthorship W3018301625A5017095669 @default.
- W3018301625 hasAuthorship W3018301625A5068081385 @default.
- W3018301625 hasAuthorship W3018301625A5083125799 @default.
- W3018301625 hasAuthorship W3018301625A5085139454 @default.
- W3018301625 hasBestOaLocation W30183016251 @default.
- W3018301625 hasConcept C108583219 @default.
- W3018301625 hasConcept C119857082 @default.
- W3018301625 hasConcept C121332964 @default.
- W3018301625 hasConcept C144024400 @default.
- W3018301625 hasConcept C153180895 @default.
- W3018301625 hasConcept C154945302 @default.
- W3018301625 hasConcept C163258240 @default.
- W3018301625 hasConcept C184297639 @default.
- W3018301625 hasConcept C2779304628 @default.
- W3018301625 hasConcept C2780992000 @default.
- W3018301625 hasConcept C31510193 @default.
- W3018301625 hasConcept C36289849 @default.
- W3018301625 hasConcept C41008148 @default.
- W3018301625 hasConcept C62520636 @default.
- W3018301625 hasConceptScore W3018301625C108583219 @default.
- W3018301625 hasConceptScore W3018301625C119857082 @default.
- W3018301625 hasConceptScore W3018301625C121332964 @default.
- W3018301625 hasConceptScore W3018301625C144024400 @default.
- W3018301625 hasConceptScore W3018301625C153180895 @default.
- W3018301625 hasConceptScore W3018301625C154945302 @default.
- W3018301625 hasConceptScore W3018301625C163258240 @default.
- W3018301625 hasConceptScore W3018301625C184297639 @default.
- W3018301625 hasConceptScore W3018301625C2779304628 @default.
- W3018301625 hasConceptScore W3018301625C2780992000 @default.
- W3018301625 hasConceptScore W3018301625C31510193 @default.
- W3018301625 hasConceptScore W3018301625C36289849 @default.
- W3018301625 hasConceptScore W3018301625C41008148 @default.
- W3018301625 hasConceptScore W3018301625C62520636 @default.
- W3018301625 hasLocation W30183016251 @default.
- W3018301625 hasLocation W30183016252 @default.
- W3018301625 hasLocation W30183016253 @default.
- W3018301625 hasLocation W30183016254 @default.
- W3018301625 hasOpenAccess W3018301625 @default.
- W3018301625 hasPrimaryLocation W30183016251 @default.
- W3018301625 hasRelatedWork W1775397219 @default.
- W3018301625 hasRelatedWork W2347601237 @default.
- W3018301625 hasRelatedWork W2897995864 @default.
- W3018301625 hasRelatedWork W3086983552 @default.
- W3018301625 hasRelatedWork W4223943233 @default.
- W3018301625 hasRelatedWork W4312200629 @default.
- W3018301625 hasRelatedWork W4360585206 @default.
- W3018301625 hasRelatedWork W4364306694 @default.
- W3018301625 hasRelatedWork W4380075502 @default.
- W3018301625 hasRelatedWork W4380086463 @default.
- W3018301625 hasVolume "305" @default.
- W3018301625 isParatext "false" @default.
- W3018301625 isRetracted "false" @default.
- W3018301625 magId "3018301625" @default.