Matches in SemOpenAlex for { <https://semopenalex.org/work/W4225582783> ?p ?o ?g. }
Showing items 1 to 79 of
79
with 100 items per page.
- W4225582783 abstract "Neural network architectures have been extensively employed in the fair representation learning setting, where the objective is to learn a new representation for a given vector which is independent of sensitive information. Various representation debiasing techniques have been proposed in the literature. However, as neural networks are inherently opaque, these methods are hard to comprehend, which limits their usefulness. We propose a new framework for fair representation learning that is centered around the learning of correction vectors, which have the same dimensionality as the given data vectors. Correction vectors may be computed either explicitly via architectural constraints or implicitly by training an invertible model based on Normalizing Flows. We show experimentally that several fair representation learning models constrained in such a way do not exhibit losses in ranking or classification performance. Furthermore, we demonstrate that state-of-the-art results can be achieved by the invertible model. Finally, we discuss the law standing of our methodology in light of recent legislation in the European Union." @default.
- W4225582783 created "2022-05-05" @default.
- W4225582783 creator A5004944606 @default.
- W4225582783 creator A5023570168 @default.
- W4225582783 creator A5042265681 @default.
- W4225582783 creator A5071595380 @default.
- W4225582783 creator A5079960667 @default.
- W4225582783 creator A5083690039 @default.
- W4225582783 date "2022-02-07" @default.
- W4225582783 modified "2023-09-27" @default.
- W4225582783 title "Fair Interpretable Representation Learning with Correction Vectors" @default.
- W4225582783 doi "https://doi.org/10.48550/arxiv.2202.03078" @default.
- W4225582783 hasPublicationYear "2022" @default.
- W4225582783 type Work @default.
- W4225582783 citedByCount "0" @default.
- W4225582783 crossrefType "posted-content" @default.
- W4225582783 hasAuthorship W4225582783A5004944606 @default.
- W4225582783 hasAuthorship W4225582783A5023570168 @default.
- W4225582783 hasAuthorship W4225582783A5042265681 @default.
- W4225582783 hasAuthorship W4225582783A5071595380 @default.
- W4225582783 hasAuthorship W4225582783A5079960667 @default.
- W4225582783 hasAuthorship W4225582783A5083690039 @default.
- W4225582783 hasBestOaLocation W42255827831 @default.
- W4225582783 hasConcept C111030470 @default.
- W4225582783 hasConcept C114614502 @default.
- W4225582783 hasConcept C119857082 @default.
- W4225582783 hasConcept C154945302 @default.
- W4225582783 hasConcept C15744967 @default.
- W4225582783 hasConcept C164226766 @default.
- W4225582783 hasConcept C17744445 @default.
- W4225582783 hasConcept C188147891 @default.
- W4225582783 hasConcept C189430467 @default.
- W4225582783 hasConcept C199539241 @default.
- W4225582783 hasConcept C202444582 @default.
- W4225582783 hasConcept C2776359362 @default.
- W4225582783 hasConcept C2779458634 @default.
- W4225582783 hasConcept C33923547 @default.
- W4225582783 hasConcept C41008148 @default.
- W4225582783 hasConcept C50644808 @default.
- W4225582783 hasConcept C59404180 @default.
- W4225582783 hasConcept C80444323 @default.
- W4225582783 hasConcept C94625758 @default.
- W4225582783 hasConcept C96442724 @default.
- W4225582783 hasConceptScore W4225582783C111030470 @default.
- W4225582783 hasConceptScore W4225582783C114614502 @default.
- W4225582783 hasConceptScore W4225582783C119857082 @default.
- W4225582783 hasConceptScore W4225582783C154945302 @default.
- W4225582783 hasConceptScore W4225582783C15744967 @default.
- W4225582783 hasConceptScore W4225582783C164226766 @default.
- W4225582783 hasConceptScore W4225582783C17744445 @default.
- W4225582783 hasConceptScore W4225582783C188147891 @default.
- W4225582783 hasConceptScore W4225582783C189430467 @default.
- W4225582783 hasConceptScore W4225582783C199539241 @default.
- W4225582783 hasConceptScore W4225582783C202444582 @default.
- W4225582783 hasConceptScore W4225582783C2776359362 @default.
- W4225582783 hasConceptScore W4225582783C2779458634 @default.
- W4225582783 hasConceptScore W4225582783C33923547 @default.
- W4225582783 hasConceptScore W4225582783C41008148 @default.
- W4225582783 hasConceptScore W4225582783C50644808 @default.
- W4225582783 hasConceptScore W4225582783C59404180 @default.
- W4225582783 hasConceptScore W4225582783C80444323 @default.
- W4225582783 hasConceptScore W4225582783C94625758 @default.
- W4225582783 hasConceptScore W4225582783C96442724 @default.
- W4225582783 hasLocation W42255827831 @default.
- W4225582783 hasOpenAccess W4225582783 @default.
- W4225582783 hasPrimaryLocation W42255827831 @default.
- W4225582783 hasRelatedWork W11104910 @default.
- W4225582783 hasRelatedWork W13303319 @default.
- W4225582783 hasRelatedWork W1608636 @default.
- W4225582783 hasRelatedWork W1674447 @default.
- W4225582783 hasRelatedWork W3491078 @default.
- W4225582783 hasRelatedWork W3820303 @default.
- W4225582783 hasRelatedWork W5518722 @default.
- W4225582783 hasRelatedWork W9321062 @default.
- W4225582783 hasRelatedWork W9711757 @default.
- W4225582783 hasRelatedWork W9770290 @default.
- W4225582783 isParatext "false" @default.
- W4225582783 isRetracted "false" @default.
- W4225582783 workType "article" @default.