Matches in SemOpenAlex for { <https://semopenalex.org/work/W3201929000> ?p ?o ?g. }
Showing items 1 to 85 of
85
with 100 items per page.
- W3201929000 abstract "In this work we examine how fine-tuning impacts the fairness of contrastive Self-Supervised Learning (SSL) models. Our findings indicate that Batch Normalization (BN) statistics play a crucial role, and that updating only the BN statistics of a pre-trained SSL backbone improves its downstream fairness (36% worst subgroup, 25% mean subgroup gap). This procedure is competitive with supervised learning, while taking 4.4x less time to train and requiring only 0.35% as many parameters to be updated. Finally, inspired by recent work in supervised learning, we find that updating BN statistics and training residual skip connections (12.3% of the parameters) achieves parity with a fully fine-tuned model, while taking 1.33x less time to train." @default.
- W3201929000 created "2021-10-11" @default.
- W3201929000 creator A5020604198 @default.
- W3201929000 creator A5033975914 @default.
- W3201929000 creator A5050567760 @default.
- W3201929000 date "2021-10-01" @default.
- W3201929000 modified "2023-09-26" @default.
- W3201929000 title "Evaluating the fairness of fine-tuning strategies in self-supervised learning." @default.
- W3201929000 cites W2108598243 @default.
- W3201929000 cites W2194775991 @default.
- W3201929000 cites W2524301210 @default.
- W3201929000 cites W2622263826 @default.
- W3201929000 cites W2896457183 @default.
- W3201929000 cites W2949117887 @default.
- W3201929000 cites W2963587345 @default.
- W3201929000 cites W2964121744 @default.
- W3201929000 cites W3002868443 @default.
- W3201929000 cites W3034978746 @default.
- W3201929000 cites W3095121901 @default.
- W3201929000 cites W3100177202 @default.
- W3201929000 cites W3101821705 @default.
- W3201929000 cites W3119980997 @default.
- W3201929000 cites W3134652006 @default.
- W3201929000 cites W3159481202 @default.
- W3201929000 cites W3163602117 @default.
- W3201929000 cites W3175277764 @default.
- W3201929000 hasPublicationYear "2021" @default.
- W3201929000 type Work @default.
- W3201929000 sameAs 3201929000 @default.
- W3201929000 citedByCount "0" @default.
- W3201929000 crossrefType "posted-content" @default.
- W3201929000 hasAuthorship W3201929000A5020604198 @default.
- W3201929000 hasAuthorship W3201929000A5033975914 @default.
- W3201929000 hasAuthorship W3201929000A5050567760 @default.
- W3201929000 hasConcept C105795698 @default.
- W3201929000 hasConcept C11413529 @default.
- W3201929000 hasConcept C119857082 @default.
- W3201929000 hasConcept C136389625 @default.
- W3201929000 hasConcept C136886441 @default.
- W3201929000 hasConcept C144024400 @default.
- W3201929000 hasConcept C154945302 @default.
- W3201929000 hasConcept C155512373 @default.
- W3201929000 hasConcept C19165224 @default.
- W3201929000 hasConcept C33923547 @default.
- W3201929000 hasConcept C41008148 @default.
- W3201929000 hasConcept C50644808 @default.
- W3201929000 hasConceptScore W3201929000C105795698 @default.
- W3201929000 hasConceptScore W3201929000C11413529 @default.
- W3201929000 hasConceptScore W3201929000C119857082 @default.
- W3201929000 hasConceptScore W3201929000C136389625 @default.
- W3201929000 hasConceptScore W3201929000C136886441 @default.
- W3201929000 hasConceptScore W3201929000C144024400 @default.
- W3201929000 hasConceptScore W3201929000C154945302 @default.
- W3201929000 hasConceptScore W3201929000C155512373 @default.
- W3201929000 hasConceptScore W3201929000C19165224 @default.
- W3201929000 hasConceptScore W3201929000C33923547 @default.
- W3201929000 hasConceptScore W3201929000C41008148 @default.
- W3201929000 hasConceptScore W3201929000C50644808 @default.
- W3201929000 hasLocation W32019290001 @default.
- W3201929000 hasOpenAccess W3201929000 @default.
- W3201929000 hasPrimaryLocation W32019290001 @default.
- W3201929000 hasRelatedWork W2435367904 @default.
- W3201929000 hasRelatedWork W2564034046 @default.
- W3201929000 hasRelatedWork W2604921058 @default.
- W3201929000 hasRelatedWork W2765297602 @default.
- W3201929000 hasRelatedWork W2911555067 @default.
- W3201929000 hasRelatedWork W3007031922 @default.
- W3201929000 hasRelatedWork W3034269976 @default.
- W3201929000 hasRelatedWork W3035324854 @default.
- W3201929000 hasRelatedWork W3037828233 @default.
- W3201929000 hasRelatedWork W3126916000 @default.
- W3201929000 hasRelatedWork W3129049565 @default.
- W3201929000 hasRelatedWork W3131944163 @default.
- W3201929000 hasRelatedWork W3148342934 @default.
- W3201929000 hasRelatedWork W3165956705 @default.
- W3201929000 hasRelatedWork W3169363132 @default.
- W3201929000 hasRelatedWork W3199451663 @default.
- W3201929000 hasRelatedWork W3202439292 @default.
- W3201929000 hasRelatedWork W3205794883 @default.
- W3201929000 hasRelatedWork W3211540990 @default.
- W3201929000 hasRelatedWork W3212159807 @default.
- W3201929000 isParatext "false" @default.
- W3201929000 isRetracted "false" @default.
- W3201929000 magId "3201929000" @default.
- W3201929000 workType "article" @default.