Matches in SemOpenAlex for { <https://semopenalex.org/work/W3213159541> ?p ?o ?g. }
Showing items 1 to 82 of
82
with 100 items per page.
- W3213159541 abstract "Question Answering (QA) tasks requiring information from multiple documents often rely on a retrieval model to identify relevant information for reasoning. The retrieval model is typically trained to maximize the likelihood of the labeled supporting evidence. However, when retrieving from large text corpora such as Wikipedia, the correct answer can often be obtained from multiple evidence candidates. Moreover, not all such candidates are labeled as positive during annotation, rendering the training signal weak and noisy. This problem is exacerbated when the questions are unanswerable or when the answers are Boolean, since the model cannot rely on lexical overlap to make a connection between the answer and supporting evidence. We develop a new parameterization of set-valued retrieval that handles unanswerable queries, and we show that marginalizing over this set during training allows a model to mitigate false negatives in supporting evidence annotations. We test our method on two multi-document QA datasets, IIRC and HotpotQA. On IIRC, we show that joint modeling with marginalization improves model performance by 5.5 F1 points and achieves a new state-of-the-art performance of 50.5 F1. We also show that retrieval marginalization results in 4.1 QA F1 improvement over a non-marginalized baseline on HotpotQA in the fullwiki setting." @default.
- W3213159541 created "2021-11-22" @default.
- W3213159541 creator A5029074038 @default.
- W3213159541 creator A5035088083 @default.
- W3213159541 creator A5087977697 @default.
- W3213159541 date "2021-01-01" @default.
- W3213159541 modified "2023-10-18" @default.
- W3213159541 title "Mitigating False-Negative Contexts in Multi-document Question Answering with Retrieval Marginalization" @default.
- W3213159541 cites W2252136820 @default.
- W3213159541 cites W2757361303 @default.
- W3213159541 cites W2889787757 @default.
- W3213159541 cites W2912924812 @default.
- W3213159541 cites W2949847757 @default.
- W3213159541 cites W2951434086 @default.
- W3213159541 cites W2962718483 @default.
- W3213159541 cites W2963341956 @default.
- W3213159541 cites W2970139579 @default.
- W3213159541 cites W2970900584 @default.
- W3213159541 cites W2995638926 @default.
- W3213159541 cites W2995649781 @default.
- W3213159541 cites W2996035354 @default.
- W3213159541 cites W3003186568 @default.
- W3213159541 cites W3027879771 @default.
- W3213159541 cites W3045462440 @default.
- W3213159541 cites W3099524945 @default.
- W3213159541 cites W3099700870 @default.
- W3213159541 cites W3099876468 @default.
- W3213159541 cites W3105055324 @default.
- W3213159541 cites W3182778088 @default.
- W3213159541 doi "https://doi.org/10.18653/v1/2021.emnlp-main.497" @default.
- W3213159541 hasPublicationYear "2021" @default.
- W3213159541 type Work @default.
- W3213159541 sameAs 3213159541 @default.
- W3213159541 citedByCount "1" @default.
- W3213159541 countsByYear W32131595412022 @default.
- W3213159541 crossrefType "proceedings-article" @default.
- W3213159541 hasAuthorship W3213159541A5029074038 @default.
- W3213159541 hasAuthorship W3213159541A5035088083 @default.
- W3213159541 hasAuthorship W3213159541A5087977697 @default.
- W3213159541 hasBestOaLocation W32131595411 @default.
- W3213159541 hasConcept C111368507 @default.
- W3213159541 hasConcept C12725497 @default.
- W3213159541 hasConcept C127313418 @default.
- W3213159541 hasConcept C154945302 @default.
- W3213159541 hasConcept C177264268 @default.
- W3213159541 hasConcept C199360897 @default.
- W3213159541 hasConcept C205711294 @default.
- W3213159541 hasConcept C23123220 @default.
- W3213159541 hasConcept C2776321320 @default.
- W3213159541 hasConcept C41008148 @default.
- W3213159541 hasConcept C44291984 @default.
- W3213159541 hasConcept C51632099 @default.
- W3213159541 hasConceptScore W3213159541C111368507 @default.
- W3213159541 hasConceptScore W3213159541C12725497 @default.
- W3213159541 hasConceptScore W3213159541C127313418 @default.
- W3213159541 hasConceptScore W3213159541C154945302 @default.
- W3213159541 hasConceptScore W3213159541C177264268 @default.
- W3213159541 hasConceptScore W3213159541C199360897 @default.
- W3213159541 hasConceptScore W3213159541C205711294 @default.
- W3213159541 hasConceptScore W3213159541C23123220 @default.
- W3213159541 hasConceptScore W3213159541C2776321320 @default.
- W3213159541 hasConceptScore W3213159541C41008148 @default.
- W3213159541 hasConceptScore W3213159541C44291984 @default.
- W3213159541 hasConceptScore W3213159541C51632099 @default.
- W3213159541 hasLocation W32131595411 @default.
- W3213159541 hasLocation W32131595412 @default.
- W3213159541 hasOpenAccess W3213159541 @default.
- W3213159541 hasPrimaryLocation W32131595411 @default.
- W3213159541 hasRelatedWork W11196620 @default.
- W3213159541 hasRelatedWork W11973946 @default.
- W3213159541 hasRelatedWork W13249535 @default.
- W3213159541 hasRelatedWork W13779650 @default.
- W3213159541 hasRelatedWork W14310100 @default.
- W3213159541 hasRelatedWork W14942459 @default.
- W3213159541 hasRelatedWork W4135815 @default.
- W3213159541 hasRelatedWork W5516682 @default.
- W3213159541 hasRelatedWork W6865571 @default.
- W3213159541 hasRelatedWork W9224363 @default.
- W3213159541 isParatext "false" @default.
- W3213159541 isRetracted "false" @default.
- W3213159541 magId "3213159541" @default.
- W3213159541 workType "article" @default.