FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3213681714> ?p ?o ?g. }

• W3213681714 abstract "Among the most critical limitations of deep learning NLP models are their lack of interpretability, and their reliance on spurious correlations. Prior work proposed various approaches to interpreting the black-box models to unveil the spurious correlations, but the research was primarily used in human-computer interaction scenarios. It still remains underexplored whether or how such model interpretations can be used to automatically “unlearn” confounding features. In this work, we propose influence tuning—a procedure that leverages model interpretations to update the model parameters towards a plausible interpretation (rather than an interpretation that relies on spurious patterns in the data) in addition to learning to predict the task labels. We show that in a controlled setup, influence tuning can help deconfounding the model from spurious patterns in data, significantly outperforming baseline methods that use adversarial training." @default.
• W3213681714 created "2021-11-22" @default.
• W3213681714 creator A5062910836 @default.
• W3213681714 creator A5091822496 @default.
• W3213681714 date "2021-01-01" @default.
• W3213681714 modified "2023-10-16" @default.
• W3213681714 title "Influence Tuning: Demoting Spurious Correlations via Instance Attribution and Instance-Driven Updates" @default.
• W3213681714 cites W1731081199 @default.
• W3213681714 cites W2109378394 @default.
• W3213681714 cites W2282821441 @default.
• W3213681714 cites W2597603852 @default.
• W3213681714 cites W2798665661 @default.
• W3213681714 cites W2802105481 @default.
• W3213681714 cites W2804178350 @default.
• W3213681714 cites W2888161220 @default.
• W3213681714 cites W2890457888 @default.
• W3213681714 cites W2899682268 @default.
• W3213681714 cites W2905678540 @default.
• W3213681714 cites W2906152891 @default.
• W3213681714 cites W2934842096 @default.
• W3213681714 cites W2949678053 @default.
• W3213681714 cites W2950866572 @default.
• W3213681714 cites W2950888501 @default.
• W3213681714 cites W2951286828 @default.
• W3213681714 cites W2962736243 @default.
• W3213681714 cites W2962772482 @default.
• W3213681714 cites W2962790223 @default.
• W3213681714 cites W2962851944 @default.
• W3213681714 cites W2963341956 @default.
• W3213681714 cites W2963395533 @default.
• W3213681714 cites W2963446520 @default.
• W3213681714 cites W2963798744 @default.
• W3213681714 cites W2963969878 @default.
• W3213681714 cites W2964159778 @default.
• W3213681714 cites W2966924523 @default.
• W3213681714 cites W2970019270 @default.
• W3213681714 cites W2970191075 @default.
• W3213681714 cites W2970726176 @default.
• W3213681714 cites W2991497298 @default.
• W3213681714 cites W2994934025 @default.
• W3213681714 cites W3034475796 @default.
• W3213681714 cites W3037831233 @default.
• W3213681714 cites W3037881973 @default.
• W3213681714 cites W3098613713 @default.
• W3213681714 cites W3101656801 @default.
• W3213681714 cites W3114796327 @default.
• W3213681714 cites W3131061281 @default.
• W3213681714 cites W3152911627 @default.
• W3213681714 cites W3154565472 @default.
• W3213681714 cites W3157894658 @default.
• W3213681714 cites W3174785103 @default.
• W3213681714 cites W3175765954 @default.
• W3213681714 cites W3213898543 @default.
• W3213681714 doi "https://doi.org/10.18653/v1/2021.findings-emnlp.374" @default.
• W3213681714 hasPublicationYear "2021" @default.
• W3213681714 type Work @default.
• W3213681714 sameAs 3213681714 @default.
• W3213681714 citedByCount "2" @default.
• W3213681714 countsByYear W32136817142021 @default.
• W3213681714 crossrefType "proceedings-article" @default.
• W3213681714 hasAuthorship W3213681714A5062910836 @default.
• W3213681714 hasAuthorship W3213681714A5091822496 @default.
• W3213681714 hasBestOaLocation W32136817141 @default.
• W3213681714 hasConcept C119857082 @default.
• W3213681714 hasConcept C154945302 @default.
• W3213681714 hasConcept C162324750 @default.
• W3213681714 hasConcept C187736073 @default.
• W3213681714 hasConcept C199360897 @default.
• W3213681714 hasConcept C2780451532 @default.
• W3213681714 hasConcept C2781067378 @default.
• W3213681714 hasConcept C41008148 @default.
• W3213681714 hasConcept C527412718 @default.
• W3213681714 hasConcept C97256817 @default.
• W3213681714 hasConceptScore W3213681714C119857082 @default.
• W3213681714 hasConceptScore W3213681714C154945302 @default.
• W3213681714 hasConceptScore W3213681714C162324750 @default.
• W3213681714 hasConceptScore W3213681714C187736073 @default.
• W3213681714 hasConceptScore W3213681714C199360897 @default.
• W3213681714 hasConceptScore W3213681714C2780451532 @default.
• W3213681714 hasConceptScore W3213681714C2781067378 @default.
• W3213681714 hasConceptScore W3213681714C41008148 @default.
• W3213681714 hasConceptScore W3213681714C527412718 @default.
• W3213681714 hasConceptScore W3213681714C97256817 @default.
• W3213681714 hasLocation W32136817141 @default.
• W3213681714 hasLocation W32136817142 @default.
• W3213681714 hasOpenAccess W3213681714 @default.
• W3213681714 hasPrimaryLocation W32136817141 @default.
• W3213681714 hasRelatedWork W2605281151 @default.
• W3213681714 hasRelatedWork W3006943036 @default.
• W3213681714 hasRelatedWork W3012234327 @default.
• W3213681714 hasRelatedWork W3119715496 @default.
• W3213681714 hasRelatedWork W3191046242 @default.
• W3213681714 hasRelatedWork W4205364923 @default.
• W3213681714 hasRelatedWork W4206534706 @default.
• W3213681714 hasRelatedWork W4229079080 @default.
• W3213681714 hasRelatedWork W4294031299 @default.
• W3213681714 hasRelatedWork W4366769859 @default.
• W3213681714 isParatext "false" @default.
• W3213681714 isRetracted "false" @default.
• W3213681714 magId "3213681714" @default.

• next