Matches in SemOpenAlex for { <https://semopenalex.org/work/W2989430018> ?p ?o ?g. }
- W2989430018 abstract "The creation of large-scale open domain reading comprehension data sets in recent years has enabled the development of end-to-end neural comprehension models with promising results. To use these models for domains with limited training data, one of the most effective approach is to first pretrain them on large out-of-domain source data and then fine-tune them with the limited target data. The caveat of this is that after fine-tuning the comprehension models tend to perform poorly in the source domain, a phenomenon known as catastrophic forgetting. In this paper, we explore methods that overcome catastrophic forgetting during fine-tuning without assuming access to data from the source domain. We introduce new auxiliary penalty terms and observe the best performance when a combination of auxiliary penalty terms is used to regularise the fine-tuning process for adapting comprehension models. To test our methods, we develop and release 6 narrow domain data sets that could potentially be used as reading comprehension benchmarks." @default.
- W2989430018 created "2019-11-22" @default.
- W2989430018 creator A5006844787 @default.
- W2989430018 creator A5032767467 @default.
- W2989430018 creator A5037751331 @default.
- W2989430018 creator A5040147917 @default.
- W2989430018 date "2019-11-01" @default.
- W2989430018 modified "2023-10-18" @default.
- W2989430018 title "Forget Me Not: Reducing Catastrophic Forgetting for Domain Adaptation in Reading Comprehension" @default.
- W2989430018 cites W1682403713 @default.
- W2989430018 cites W1880262756 @default.
- W2989430018 cites W1981208470 @default.
- W2989430018 cites W2060277733 @default.
- W2989430018 cites W2120354757 @default.
- W2989430018 cites W2123442489 @default.
- W2989430018 cites W2250539671 @default.
- W2989430018 cites W2251818205 @default.
- W2989430018 cites W2523586319 @default.
- W2989430018 cites W2540485556 @default.
- W2989430018 cites W2558203065 @default.
- W2989430018 cites W2560647685 @default.
- W2989430018 cites W2612017161 @default.
- W2989430018 cites W2782276075 @default.
- W2989430018 cites W2809324505 @default.
- W2989430018 cites W2898290017 @default.
- W2989430018 cites W2912924812 @default.
- W2989430018 cites W2951534261 @default.
- W2989430018 cites W2962724315 @default.
- W2989430018 cites W2963323070 @default.
- W2989430018 cites W2963339397 @default.
- W2989430018 cites W2963341956 @default.
- W2989430018 cites W2963403868 @default.
- W2989430018 cites W2963488798 @default.
- W2989430018 cites W2963564796 @default.
- W2989430018 cites W2963748441 @default.
- W2989430018 cites W2963876099 @default.
- W2989430018 cites W2963963993 @default.
- W2989430018 doi "https://doi.org/10.48550/arxiv.1911.00202" @default.
- W2989430018 hasPublicationYear "2019" @default.
- W2989430018 type Work @default.
- W2989430018 sameAs 2989430018 @default.
- W2989430018 citedByCount "1" @default.
- W2989430018 countsByYear W29894300182020 @default.
- W2989430018 crossrefType "posted-content" @default.
- W2989430018 hasAuthorship W2989430018A5006844787 @default.
- W2989430018 hasAuthorship W2989430018A5032767467 @default.
- W2989430018 hasAuthorship W2989430018A5037751331 @default.
- W2989430018 hasAuthorship W2989430018A5040147917 @default.
- W2989430018 hasBestOaLocation W29894300181 @default.
- W2989430018 hasConcept C115903868 @default.
- W2989430018 hasConcept C134306372 @default.
- W2989430018 hasConcept C138885662 @default.
- W2989430018 hasConcept C149091818 @default.
- W2989430018 hasConcept C154945302 @default.
- W2989430018 hasConcept C15744967 @default.
- W2989430018 hasConcept C16910744 @default.
- W2989430018 hasConcept C180747234 @default.
- W2989430018 hasConcept C199360897 @default.
- W2989430018 hasConcept C204321447 @default.
- W2989430018 hasConcept C2776434776 @default.
- W2989430018 hasConcept C2777561058 @default.
- W2989430018 hasConcept C2777904410 @default.
- W2989430018 hasConcept C2778780117 @default.
- W2989430018 hasConcept C33923547 @default.
- W2989430018 hasConcept C36503486 @default.
- W2989430018 hasConcept C41008148 @default.
- W2989430018 hasConcept C41895202 @default.
- W2989430018 hasConcept C511192102 @default.
- W2989430018 hasConcept C554936623 @default.
- W2989430018 hasConcept C7149132 @default.
- W2989430018 hasConcept C98045186 @default.
- W2989430018 hasConceptScore W2989430018C115903868 @default.
- W2989430018 hasConceptScore W2989430018C134306372 @default.
- W2989430018 hasConceptScore W2989430018C138885662 @default.
- W2989430018 hasConceptScore W2989430018C149091818 @default.
- W2989430018 hasConceptScore W2989430018C154945302 @default.
- W2989430018 hasConceptScore W2989430018C15744967 @default.
- W2989430018 hasConceptScore W2989430018C16910744 @default.
- W2989430018 hasConceptScore W2989430018C180747234 @default.
- W2989430018 hasConceptScore W2989430018C199360897 @default.
- W2989430018 hasConceptScore W2989430018C204321447 @default.
- W2989430018 hasConceptScore W2989430018C2776434776 @default.
- W2989430018 hasConceptScore W2989430018C2777561058 @default.
- W2989430018 hasConceptScore W2989430018C2777904410 @default.
- W2989430018 hasConceptScore W2989430018C2778780117 @default.
- W2989430018 hasConceptScore W2989430018C33923547 @default.
- W2989430018 hasConceptScore W2989430018C36503486 @default.
- W2989430018 hasConceptScore W2989430018C41008148 @default.
- W2989430018 hasConceptScore W2989430018C41895202 @default.
- W2989430018 hasConceptScore W2989430018C511192102 @default.
- W2989430018 hasConceptScore W2989430018C554936623 @default.
- W2989430018 hasConceptScore W2989430018C7149132 @default.
- W2989430018 hasConceptScore W2989430018C98045186 @default.
- W2989430018 hasLocation W29894300181 @default.
- W2989430018 hasOpenAccess W2989430018 @default.
- W2989430018 hasPrimaryLocation W29894300181 @default.
- W2989430018 hasRelatedWork W1864954421 @default.
- W2989430018 hasRelatedWork W1969359128 @default.
- W2989430018 hasRelatedWork W1977813563 @default.
- W2989430018 hasRelatedWork W1982811735 @default.