Matches in SemOpenAlex for { <https://semopenalex.org/work/W3106096977> ?p ?o ?g. }
Showing items 1 to 88 of
88
with 100 items per page.
- W3106096977 abstract "Existing NLP datasets contain various biases that models can easily exploit to achieve high performances on the corresponding evaluation sets. However, focusing on dataset-specific biases limits their ability to learn more generalizable knowledge about the task from more general data patterns. In this paper, we investigate the impact of debiasing methods for improving generalization and propose a general framework for improving the performance on both in-domain and out-of-domain datasets by concurrent modeling of multiple biases in the training data. Our framework weights each example based on the biases it contains and the strength of those biases in the training data. It then uses these weights in the training objective so that the model relies less on examples with high bias weights. We extensively evaluate our framework on extractive question answering with training data from various domains with multiple biases of different strengths. We perform the evaluations in two different settings, in which the model is trained on a single domain or multiple domains simultaneously, and show its effectiveness in both settings compared to state-of-the-art debiasing methods." @default.
- W3106096977 created "2020-11-23" @default.
- W3106096977 creator A5014688346 @default.
- W3106096977 creator A5015236807 @default.
- W3106096977 creator A5027450194 @default.
- W3106096977 creator A5054918343 @default.
- W3106096977 date "2020-01-01" @default.
- W3106096977 modified "2023-10-02" @default.
- W3106096977 title "Improving QA Generalization by Concurrent Modeling of Multiple Biases" @default.
- W3106096977 cites W1821462560 @default.
- W3106096977 cites W2551396370 @default.
- W3106096977 cites W2557764419 @default.
- W3106096977 cites W2606964149 @default.
- W3106096977 cites W2626154462 @default.
- W3106096977 cites W2746097825 @default.
- W3106096977 cites W2889787757 @default.
- W3106096977 cites W2912924812 @default.
- W3106096977 cites W2919420119 @default.
- W3106096977 cites W2951286828 @default.
- W3106096977 cites W2951873305 @default.
- W3106096977 cites W2962736243 @default.
- W3106096977 cites W2962881743 @default.
- W3106096977 cites W2963339397 @default.
- W3106096977 cites W2963341956 @default.
- W3106096977 cites W2963542100 @default.
- W3106096977 cites W2963691697 @default.
- W3106096977 cites W2963748441 @default.
- W3106096977 cites W2963796403 @default.
- W3106096977 cites W2963928014 @default.
- W3106096977 cites W2963969878 @default.
- W3106096977 cites W2964044490 @default.
- W3106096977 cites W2970019270 @default.
- W3106096977 cites W2970379526 @default.
- W3106096977 cites W2970597249 @default.
- W3106096977 cites W2984402309 @default.
- W3106096977 cites W2985436116 @default.
- W3106096977 cites W2986532682 @default.
- W3106096977 cites W2987241137 @default.
- W3106096977 cites W2988421999 @default.
- W3106096977 cites W2997200074 @default.
- W3106096977 cites W2998072062 @default.
- W3106096977 cites W3034831508 @default.
- W3106096977 cites W3035032873 @default.
- W3106096977 cites W3035139434 @default.
- W3106096977 cites W3103361051 @default.
- W3106096977 doi "https://doi.org/10.18653/v1/2020.findings-emnlp.74" @default.
- W3106096977 hasPublicationYear "2020" @default.
- W3106096977 type Work @default.
- W3106096977 sameAs 3106096977 @default.
- W3106096977 citedByCount "8" @default.
- W3106096977 countsByYear W31060969772020 @default.
- W3106096977 countsByYear W31060969772021 @default.
- W3106096977 crossrefType "proceedings-article" @default.
- W3106096977 hasAuthorship W3106096977A5014688346 @default.
- W3106096977 hasAuthorship W3106096977A5015236807 @default.
- W3106096977 hasAuthorship W3106096977A5027450194 @default.
- W3106096977 hasAuthorship W3106096977A5054918343 @default.
- W3106096977 hasBestOaLocation W31060969771 @default.
- W3106096977 hasConcept C134306372 @default.
- W3106096977 hasConcept C154945302 @default.
- W3106096977 hasConcept C177148314 @default.
- W3106096977 hasConcept C33923547 @default.
- W3106096977 hasConcept C41008148 @default.
- W3106096977 hasConceptScore W3106096977C134306372 @default.
- W3106096977 hasConceptScore W3106096977C154945302 @default.
- W3106096977 hasConceptScore W3106096977C177148314 @default.
- W3106096977 hasConceptScore W3106096977C33923547 @default.
- W3106096977 hasConceptScore W3106096977C41008148 @default.
- W3106096977 hasLocation W31060969771 @default.
- W3106096977 hasLocation W31060969772 @default.
- W3106096977 hasLocation W31060969773 @default.
- W3106096977 hasLocation W31060969774 @default.
- W3106096977 hasOpenAccess W3106096977 @default.
- W3106096977 hasPrimaryLocation W31060969771 @default.
- W3106096977 hasRelatedWork W2049775471 @default.
- W3106096977 hasRelatedWork W2093578348 @default.
- W3106096977 hasRelatedWork W2350741829 @default.
- W3106096977 hasRelatedWork W2358668433 @default.
- W3106096977 hasRelatedWork W2376932109 @default.
- W3106096977 hasRelatedWork W2382290278 @default.
- W3106096977 hasRelatedWork W2390279801 @default.
- W3106096977 hasRelatedWork W2748952813 @default.
- W3106096977 hasRelatedWork W2899084033 @default.
- W3106096977 hasRelatedWork W2989932438 @default.
- W3106096977 isParatext "false" @default.
- W3106096977 isRetracted "false" @default.
- W3106096977 magId "3106096977" @default.
- W3106096977 workType "article" @default.