FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3099606977> ?p ?o ?g. }

• W3099606977 abstract "While discriminative neural network classifiers are generally preferred, recent work has shown advantages of generative classifiers in term of data efficiency and robustness. In this paper, we focus on natural language inference (NLI). We propose GenNLI, a generative classifier for NLI tasks, and empirically characterize its performance by comparing it to five baselines, including discriminative models and large-scale pretrained language representation models like BERT. We explore training objectives for discriminative fine-tuning of our generative classifiers, showing improvements over log loss fine-tuning from prior work (Lewis and Fan, 2019). In particular, we find strong results with a simple unbounded modification to log loss, which we call the “infinilog loss”. Our experiments show that GenNLI outperforms both discriminative and pretrained baselines across several challenging NLI experimental settings, including small training sets, imbalanced label distributions, and label noise." @default.
• W3099606977 created "2020-11-23" @default.
• W3099606977 creator A5007956930 @default.
• W3099606977 creator A5021459300 @default.
• W3099606977 creator A5038032242 @default.
• W3099606977 creator A5077889889 @default.
• W3099606977 creator A5081022650 @default.
• W3099606977 date "2020-01-01" @default.
• W3099606977 modified "2023-10-16" @default.
• W3099606977 title "Discriminatively-Tuned Generative Classifiers for Robust Natural Language Inference" @default.
• W3099606977 cites W14913773 @default.
• W3099606977 cites W1840435438 @default.
• W3099606977 cites W1902237438 @default.
• W3099606977 cites W1980776243 @default.
• W3099606977 cites W2040870580 @default.
• W3099606977 cites W2064675550 @default.
• W3099606977 cites W2105644991 @default.
• W3099606977 cites W2109553965 @default.
• W3099606977 cites W2117130368 @default.
• W3099606977 cites W2125234026 @default.
• W3099606977 cites W2137143056 @default.
• W3099606977 cites W2154474435 @default.
• W3099606977 cites W2228826686 @default.
• W3099606977 cites W2250539671 @default.
• W3099606977 cites W2250790822 @default.
• W3099606977 cites W2413794162 @default.
• W3099606977 cites W2593887162 @default.
• W3099606977 cites W2608787653 @default.
• W3099606977 cites W2620760558 @default.
• W3099606977 cites W2798665661 @default.
• W3099606977 cites W2888491130 @default.
• W3099606977 cites W2889769646 @default.
• W3099606977 cites W2913222130 @default.
• W3099606977 cites W2951286828 @default.
• W3099606977 cites W2962685628 @default.
• W3099606977 cites W2962736243 @default.
• W3099606977 cites W2962739339 @default.
• W3099606977 cites W2962843521 @default.
• W3099606977 cites W2963310665 @default.
• W3099606977 cites W2963341956 @default.
• W3099606977 cites W2963403868 @default.
• W3099606977 cites W2963580443 @default.
• W3099606977 cites W2963719234 @default.
• W3099606977 cites W2963748441 @default.
• W3099606977 cites W2963846996 @default.
• W3099606977 cites W2963918774 @default.
• W3099606977 cites W2964072386 @default.
• W3099606977 cites W2964121744 @default.
• W3099606977 cites W2964165364 @default.
• W3099606977 cites W2965373594 @default.
• W3099606977 cites W2970597249 @default.
• W3099606977 cites W2986712448 @default.
• W3099606977 cites W3031416356 @default.
• W3099606977 cites W3101853775 @default.
• W3099606977 cites W2525127255 @default.
• W3099606977 doi "https://doi.org/10.18653/v1/2020.emnlp-main.657" @default.
• W3099606977 hasPublicationYear "2020" @default.
• W3099606977 type Work @default.
• W3099606977 sameAs 3099606977 @default.
• W3099606977 citedByCount "4" @default.
• W3099606977 countsByYear W30996069772020 @default.
• W3099606977 countsByYear W30996069772021 @default.
• W3099606977 crossrefType "proceedings-article" @default.
• W3099606977 hasAuthorship W3099606977A5007956930 @default.
• W3099606977 hasAuthorship W3099606977A5021459300 @default.
• W3099606977 hasAuthorship W3099606977A5038032242 @default.
• W3099606977 hasAuthorship W3099606977A5077889889 @default.
• W3099606977 hasAuthorship W3099606977A5081022650 @default.
• W3099606977 hasBestOaLocation W30996069771 @default.
• W3099606977 hasConcept C104317684 @default.
• W3099606977 hasConcept C119857082 @default.
• W3099606977 hasConcept C153180895 @default.
• W3099606977 hasConcept C154945302 @default.
• W3099606977 hasConcept C167966045 @default.
• W3099606977 hasConcept C185592680 @default.
• W3099606977 hasConcept C2776214188 @default.
• W3099606977 hasConcept C28490314 @default.
• W3099606977 hasConcept C39890363 @default.
• W3099606977 hasConcept C41008148 @default.
• W3099606977 hasConcept C51632099 @default.
• W3099606977 hasConcept C55493867 @default.
• W3099606977 hasConcept C63479239 @default.
• W3099606977 hasConcept C95623464 @default.
• W3099606977 hasConcept C97931131 @default.
• W3099606977 hasConceptScore W3099606977C104317684 @default.
• W3099606977 hasConceptScore W3099606977C119857082 @default.
• W3099606977 hasConceptScore W3099606977C153180895 @default.
• W3099606977 hasConceptScore W3099606977C154945302 @default.
• W3099606977 hasConceptScore W3099606977C167966045 @default.
• W3099606977 hasConceptScore W3099606977C185592680 @default.
• W3099606977 hasConceptScore W3099606977C2776214188 @default.
• W3099606977 hasConceptScore W3099606977C28490314 @default.
• W3099606977 hasConceptScore W3099606977C39890363 @default.
• W3099606977 hasConceptScore W3099606977C41008148 @default.
• W3099606977 hasConceptScore W3099606977C51632099 @default.
• W3099606977 hasConceptScore W3099606977C55493867 @default.
• W3099606977 hasConceptScore W3099606977C63479239 @default.
• W3099606977 hasConceptScore W3099606977C95623464 @default.
• W3099606977 hasConceptScore W3099606977C97931131 @default.
• W3099606977 hasLocation W30996069771 @default.

• next