Matches in SemOpenAlex for { <https://semopenalex.org/work/W4385573570> ?p ?o ?g. }
Showing items 1 to 53 of
53
with 100 items per page.
- W4385573570 abstract "Weakly supervised text classification methods typically train a deep neural classifier based on pseudo-labels. The quality of pseudo-labels is crucial to final performance but they are inevitably noisy due to their heuristic nature, so selecting the correct ones has a huge potential for performance boost. One straightforward solution is to select samples based on the softmax probability scores in the neural classifier corresponding to their pseudo-labels. However, we show through our experiments that such solutions are ineffective and unstable due to the erroneously high-confidence predictions from poorly calibrated models. Recent studies on the memorization effects of deep neural models suggest that these models first memorize training samples with clean labels and then those with noisy labels. Inspired by this observation, we propose a novel pseudo-label selection method LOPS that takes learning order of samples into consideration. We hypothesize that the learning order reflects the probability of wrong annotation in terms of ranking, and therefore, propose to select the samples that are learnt earlier. LOPS can be viewed as a strong performance-boost plug-in to most existing weakly-supervised text classification methods, as confirmed in extensive experiments on four real-world datasets." @default.
- W4385573570 created "2023-08-05" @default.
- W4385573570 creator A5027505884 @default.
- W4385573570 creator A5032931351 @default.
- W4385573570 creator A5039500313 @default.
- W4385573570 date "2022-01-01" @default.
- W4385573570 modified "2023-09-24" @default.
- W4385573570 title "LOPS: Learning Order Inspired Pseudo-Label Selection for Weakly Supervised Text Classification" @default.
- W4385573570 doi "https://doi.org/10.18653/v1/2022.findings-emnlp.360" @default.
- W4385573570 hasPublicationYear "2022" @default.
- W4385573570 type Work @default.
- W4385573570 citedByCount "0" @default.
- W4385573570 crossrefType "proceedings-article" @default.
- W4385573570 hasAuthorship W4385573570A5027505884 @default.
- W4385573570 hasAuthorship W4385573570A5032931351 @default.
- W4385573570 hasAuthorship W4385573570A5039500313 @default.
- W4385573570 hasBestOaLocation W43855735701 @default.
- W4385573570 hasConcept C119857082 @default.
- W4385573570 hasConcept C145420912 @default.
- W4385573570 hasConcept C153180895 @default.
- W4385573570 hasConcept C154945302 @default.
- W4385573570 hasConcept C188441871 @default.
- W4385573570 hasConcept C30038468 @default.
- W4385573570 hasConcept C33923547 @default.
- W4385573570 hasConcept C41008148 @default.
- W4385573570 hasConcept C50644808 @default.
- W4385573570 hasConcept C95623464 @default.
- W4385573570 hasConceptScore W4385573570C119857082 @default.
- W4385573570 hasConceptScore W4385573570C145420912 @default.
- W4385573570 hasConceptScore W4385573570C153180895 @default.
- W4385573570 hasConceptScore W4385573570C154945302 @default.
- W4385573570 hasConceptScore W4385573570C188441871 @default.
- W4385573570 hasConceptScore W4385573570C30038468 @default.
- W4385573570 hasConceptScore W4385573570C33923547 @default.
- W4385573570 hasConceptScore W4385573570C41008148 @default.
- W4385573570 hasConceptScore W4385573570C50644808 @default.
- W4385573570 hasConceptScore W4385573570C95623464 @default.
- W4385573570 hasLocation W43855735701 @default.
- W4385573570 hasOpenAccess W4385573570 @default.
- W4385573570 hasPrimaryLocation W43855735701 @default.
- W4385573570 hasRelatedWork W2743258233 @default.
- W4385573570 hasRelatedWork W2771515600 @default.
- W4385573570 hasRelatedWork W2807311372 @default.
- W4385573570 hasRelatedWork W2900180889 @default.
- W4385573570 hasRelatedWork W2997969508 @default.
- W4385573570 hasRelatedWork W3004093983 @default.
- W4385573570 hasRelatedWork W3120400911 @default.
- W4385573570 hasRelatedWork W3208883981 @default.
- W4385573570 hasRelatedWork W4307834408 @default.
- W4385573570 hasRelatedWork W4320925816 @default.
- W4385573570 isParatext "false" @default.
- W4385573570 isRetracted "false" @default.
- W4385573570 workType "article" @default.