Matches in SemOpenAlex for { <https://semopenalex.org/work/W3202082017> ?p ?o ?g. }
- W3202082017 abstract "Weakly-supervised text classification has received much attention in recent years for it can alleviate the heavy burden of annotating massive data. Among them, keyword-driven methods are the mainstream where user-provided keywords are exploited to generate pseudo-labels for unlabeled texts. However, existing methods treat keywords independently, thus ignore the correlation among them, which should be useful if properly exploited. In this paper, we propose a novel framework called ClassKG to explore keyword-keyword correlation on keyword graph by GNN. Our framework is an iterative process. In each iteration, we first construct a keyword graph, so the task of assigning pseudo labels is transformed to annotating keyword subgraphs. To improve the annotation quality, we introduce a self-supervised task to pretrain a subgraph annotator, and then finetune it. With the pseudo labels generated by the subgraph annotator, we then train a text classifier to classify the unlabeled texts. Finally, we re-extract keywords from the classified texts. Extensive experiments on both long-text and short-text datasets show that our method substantially outperforms the existing ones." @default.
- W3202082017 created "2021-10-11" @default.
- W3202082017 creator A5016081741 @default.
- W3202082017 creator A5017862559 @default.
- W3202082017 creator A5018343158 @default.
- W3202082017 creator A5044757881 @default.
- W3202082017 creator A5058069028 @default.
- W3202082017 date "2021-01-01" @default.
- W3202082017 modified "2023-09-30" @default.
- W3202082017 title "Weakly-supervised Text Classification Based on Keyword Graph" @default.
- W3202082017 cites W1614298861 @default.
- W3202082017 cites W2061873838 @default.
- W3202082017 cites W2079057609 @default.
- W3202082017 cites W2098136027 @default.
- W3202082017 cites W2103931177 @default.
- W3202082017 cites W2108281845 @default.
- W3202082017 cites W2113459411 @default.
- W3202082017 cites W2120779048 @default.
- W3202082017 cites W2158329013 @default.
- W3202082017 cites W2285986798 @default.
- W3202082017 cites W2321533354 @default.
- W3202082017 cites W2613881683 @default.
- W3202082017 cites W2890931111 @default.
- W3202082017 cites W2908510526 @default.
- W3202082017 cites W2950037760 @default.
- W3202082017 cites W2962711740 @default.
- W3202082017 cites W2963012544 @default.
- W3202082017 cites W2963341956 @default.
- W3202082017 cites W2963413667 @default.
- W3202082017 cites W2963420272 @default.
- W3202082017 cites W2963858333 @default.
- W3202082017 cites W2964204137 @default.
- W3202082017 cites W2970200208 @default.
- W3202082017 cites W2982399380 @default.
- W3202082017 cites W2996428491 @default.
- W3202082017 cites W2996604169 @default.
- W3202082017 cites W3015468748 @default.
- W3202082017 cites W3034588688 @default.
- W3202082017 cites W3035524453 @default.
- W3202082017 cites W3080997787 @default.
- W3202082017 cites W3099152386 @default.
- W3202082017 cites W3101620381 @default.
- W3202082017 cites W3102960713 @default.
- W3202082017 cites W3106109117 @default.
- W3202082017 cites W3126928293 @default.
- W3202082017 cites W3166913490 @default.
- W3202082017 cites W3171416056 @default.
- W3202082017 cites W2962915384 @default.
- W3202082017 doi "https://doi.org/10.18653/v1/2021.emnlp-main.222" @default.
- W3202082017 hasPublicationYear "2021" @default.
- W3202082017 type Work @default.
- W3202082017 sameAs 3202082017 @default.
- W3202082017 citedByCount "11" @default.
- W3202082017 countsByYear W32020820172021 @default.
- W3202082017 countsByYear W32020820172022 @default.
- W3202082017 countsByYear W32020820172023 @default.
- W3202082017 crossrefType "proceedings-article" @default.
- W3202082017 hasAuthorship W3202082017A5016081741 @default.
- W3202082017 hasAuthorship W3202082017A5017862559 @default.
- W3202082017 hasAuthorship W3202082017A5018343158 @default.
- W3202082017 hasAuthorship W3202082017A5044757881 @default.
- W3202082017 hasAuthorship W3202082017A5058069028 @default.
- W3202082017 hasBestOaLocation W32020820171 @default.
- W3202082017 hasConcept C117220453 @default.
- W3202082017 hasConcept C132525143 @default.
- W3202082017 hasConcept C154945302 @default.
- W3202082017 hasConcept C162324750 @default.
- W3202082017 hasConcept C187736073 @default.
- W3202082017 hasConcept C199360897 @default.
- W3202082017 hasConcept C204321447 @default.
- W3202082017 hasConcept C23123220 @default.
- W3202082017 hasConcept C2524010 @default.
- W3202082017 hasConcept C2776321320 @default.
- W3202082017 hasConcept C2780288562 @default.
- W3202082017 hasConcept C2780451532 @default.
- W3202082017 hasConcept C2780801425 @default.
- W3202082017 hasConcept C2987255567 @default.
- W3202082017 hasConcept C2988412617 @default.
- W3202082017 hasConcept C33923547 @default.
- W3202082017 hasConcept C41008148 @default.
- W3202082017 hasConcept C80444323 @default.
- W3202082017 hasConcept C95623464 @default.
- W3202082017 hasConceptScore W3202082017C117220453 @default.
- W3202082017 hasConceptScore W3202082017C132525143 @default.
- W3202082017 hasConceptScore W3202082017C154945302 @default.
- W3202082017 hasConceptScore W3202082017C162324750 @default.
- W3202082017 hasConceptScore W3202082017C187736073 @default.
- W3202082017 hasConceptScore W3202082017C199360897 @default.
- W3202082017 hasConceptScore W3202082017C204321447 @default.
- W3202082017 hasConceptScore W3202082017C23123220 @default.
- W3202082017 hasConceptScore W3202082017C2524010 @default.
- W3202082017 hasConceptScore W3202082017C2776321320 @default.
- W3202082017 hasConceptScore W3202082017C2780288562 @default.
- W3202082017 hasConceptScore W3202082017C2780451532 @default.
- W3202082017 hasConceptScore W3202082017C2780801425 @default.
- W3202082017 hasConceptScore W3202082017C2987255567 @default.
- W3202082017 hasConceptScore W3202082017C2988412617 @default.
- W3202082017 hasConceptScore W3202082017C33923547 @default.
- W3202082017 hasConceptScore W3202082017C41008148 @default.
- W3202082017 hasConceptScore W3202082017C80444323 @default.