Matches in SemOpenAlex for { <https://semopenalex.org/work/W3023532425> ?p ?o ?g. }
- W3023532425 abstract "Machine Reading Comprehension (MRC) is an important testbed for evaluating models' natural language understanding (NLU) ability. There has been rapid progress in this area, with new models achieving impressive performance on various benchmarks. However, existing benchmarks only evaluate models on in-domain test sets without considering their robustness under test-time perturbations or adversarial attacks. To fill this important gap, we construct AdvRACE (Adversarial RACE), a new model-agnostic benchmark for evaluating the robustness of MRC models under four different types of adversarial attacks, including our novel distractor extraction and generation attacks. We show that state-of-the-art (SOTA) models are vulnerable to all of these attacks. We conclude that there is substantial room for building more robust MRC models and our benchmark can help motivate and measure progress in this area. We release our data and code at https://github.com/NoviScl/AdvRACE ." @default.
- W3023532425 created "2020-05-13" @default.
- W3023532425 creator A5003738065 @default.
- W3023532425 creator A5025839290 @default.
- W3023532425 creator A5051107002 @default.
- W3023532425 creator A5063089557 @default.
- W3023532425 creator A5069365665 @default.
- W3023532425 creator A5072441760 @default.
- W3023532425 date "2020-04-29" @default.
- W3023532425 modified "2023-09-24" @default.
- W3023532425 title "Benchmarking Robustness of Machine Reading Comprehension Models" @default.
- W3023532425 cites W2038721957 @default.
- W3023532425 cites W2250539671 @default.
- W3023532425 cites W2606964149 @default.
- W3023532425 cites W2777353073 @default.
- W3023532425 cites W2798665661 @default.
- W3023532425 cites W2799007037 @default.
- W3023532425 cites W2919420119 @default.
- W3023532425 cites W2949128310 @default.
- W3023532425 cites W2950733407 @default.
- W3023532425 cites W2951873305 @default.
- W3023532425 cites W2962718684 @default.
- W3023532425 cites W2962727366 @default.
- W3023532425 cites W2963126845 @default.
- W3023532425 cites W2963341956 @default.
- W3023532425 cites W2963661177 @default.
- W3023532425 cites W2963748441 @default.
- W3023532425 cites W2963783970 @default.
- W3023532425 cites W2963969878 @default.
- W3023532425 cites W2964087080 @default.
- W3023532425 cites W2965373594 @default.
- W3023532425 cites W2965595599 @default.
- W3023532425 cites W2966491090 @default.
- W3023532425 cites W2968154628 @default.
- W3023532425 cites W2970597249 @default.
- W3023532425 cites W2971274815 @default.
- W3023532425 cites W2981320891 @default.
- W3023532425 cites W2982756474 @default.
- W3023532425 cites W2996428491 @default.
- W3023532425 cites W2997924070 @default.
- W3023532425 cites W2998196817 @default.
- W3023532425 cites W3017462938 @default.
- W3023532425 cites W3021815119 @default.
- W3023532425 cites W3034850762 @default.
- W3023532425 cites W3035507081 @default.
- W3023532425 cites W3041064305 @default.
- W3023532425 cites W3106031450 @default.
- W3023532425 doi "https://doi.org/10.48550/arxiv.2004.14004" @default.
- W3023532425 hasPublicationYear "2020" @default.
- W3023532425 type Work @default.
- W3023532425 sameAs 3023532425 @default.
- W3023532425 citedByCount "13" @default.
- W3023532425 countsByYear W30235324252020 @default.
- W3023532425 countsByYear W30235324252021 @default.
- W3023532425 crossrefType "posted-content" @default.
- W3023532425 hasAuthorship W3023532425A5003738065 @default.
- W3023532425 hasAuthorship W3023532425A5025839290 @default.
- W3023532425 hasAuthorship W3023532425A5051107002 @default.
- W3023532425 hasAuthorship W3023532425A5063089557 @default.
- W3023532425 hasAuthorship W3023532425A5069365665 @default.
- W3023532425 hasAuthorship W3023532425A5072441760 @default.
- W3023532425 hasBestOaLocation W30235324251 @default.
- W3023532425 hasConcept C104317684 @default.
- W3023532425 hasConcept C119857082 @default.
- W3023532425 hasConcept C13280743 @default.
- W3023532425 hasConcept C144133560 @default.
- W3023532425 hasConcept C154945302 @default.
- W3023532425 hasConcept C162853370 @default.
- W3023532425 hasConcept C185592680 @default.
- W3023532425 hasConcept C185798385 @default.
- W3023532425 hasConcept C199360897 @default.
- W3023532425 hasConcept C205649164 @default.
- W3023532425 hasConcept C31258907 @default.
- W3023532425 hasConcept C31395832 @default.
- W3023532425 hasConcept C37736160 @default.
- W3023532425 hasConcept C41008148 @default.
- W3023532425 hasConcept C43126263 @default.
- W3023532425 hasConcept C55493867 @default.
- W3023532425 hasConcept C63479239 @default.
- W3023532425 hasConcept C86251818 @default.
- W3023532425 hasConceptScore W3023532425C104317684 @default.
- W3023532425 hasConceptScore W3023532425C119857082 @default.
- W3023532425 hasConceptScore W3023532425C13280743 @default.
- W3023532425 hasConceptScore W3023532425C144133560 @default.
- W3023532425 hasConceptScore W3023532425C154945302 @default.
- W3023532425 hasConceptScore W3023532425C162853370 @default.
- W3023532425 hasConceptScore W3023532425C185592680 @default.
- W3023532425 hasConceptScore W3023532425C185798385 @default.
- W3023532425 hasConceptScore W3023532425C199360897 @default.
- W3023532425 hasConceptScore W3023532425C205649164 @default.
- W3023532425 hasConceptScore W3023532425C31258907 @default.
- W3023532425 hasConceptScore W3023532425C31395832 @default.
- W3023532425 hasConceptScore W3023532425C37736160 @default.
- W3023532425 hasConceptScore W3023532425C41008148 @default.
- W3023532425 hasConceptScore W3023532425C43126263 @default.
- W3023532425 hasConceptScore W3023532425C55493867 @default.
- W3023532425 hasConceptScore W3023532425C63479239 @default.
- W3023532425 hasConceptScore W3023532425C86251818 @default.
- W3023532425 hasLocation W30235324251 @default.
- W3023532425 hasOpenAccess W3023532425 @default.