Matches in SemOpenAlex for { <https://semopenalex.org/work/W3203374547> ?p ?o ?g. }
- W3203374547 abstract "The performance of many intent detection approaches will degrade when they meet open-set data because there is out-of-domain (OOD) noise. Some works utilize clean but expensive labeled data to supervise models more robust to the varied environment. However, a large number of labeled samples are scarce and their models no longer change after finishing training. To address this problem, we propose an iterative learning framework that can dynamically improve the model's ability of OOD intent detection and meanwhile continually obtain valuable new data to learn deep discriminative features. Concretely, the model can generate pseudo labels for unlabeled examples by self-training and use the local outlier factor (LOF) algorithm to detect unknown intents. Furthermore, we add mask operation on supervised contrastive loss (SCL) and use the masked-SCL to absorb new data effectively. Experiments on CLINC and SNIPS demonstrate that our proposed method can robustly realize intent detection in the presence of a high proportion of open-set." @default.
- W3203374547 created "2021-10-11" @default.
- W3203374547 creator A5016651990 @default.
- W3203374547 creator A5027438317 @default.
- W3203374547 creator A5030878947 @default.
- W3203374547 creator A5061529715 @default.
- W3203374547 creator A5084983727 @default.
- W3203374547 creator A5085963057 @default.
- W3203374547 date "2021-07-18" @default.
- W3203374547 modified "2023-10-16" @default.
- W3203374547 title "Self-training with Masked Supervised Contrastive Loss for Unknown Intents Detection" @default.
- W3203374547 cites W1994550352 @default.
- W3203374547 cites W2088622183 @default.
- W3203374547 cites W2101210369 @default.
- W3203374547 cites W2111316763 @default.
- W3203374547 cites W2138621090 @default.
- W3203374547 cites W2144182447 @default.
- W3203374547 cites W2157364932 @default.
- W3203374547 cites W2250539671 @default.
- W3203374547 cites W2618530766 @default.
- W3203374547 cites W2803392141 @default.
- W3203374547 cites W2804945011 @default.
- W3203374547 cites W2952409498 @default.
- W3203374547 cites W2962898354 @default.
- W3203374547 cites W2963466847 @default.
- W3203374547 cites W2964155802 @default.
- W3203374547 cites W2969394406 @default.
- W3203374547 cites W2984353870 @default.
- W3203374547 cites W2986193249 @default.
- W3203374547 cites W2998721586 @default.
- W3203374547 cites W3035160371 @default.
- W3203374547 doi "https://doi.org/10.1109/ijcnn52387.2021.9534224" @default.
- W3203374547 hasPublicationYear "2021" @default.
- W3203374547 type Work @default.
- W3203374547 sameAs 3203374547 @default.
- W3203374547 citedByCount "0" @default.
- W3203374547 crossrefType "proceedings-article" @default.
- W3203374547 hasAuthorship W3203374547A5016651990 @default.
- W3203374547 hasAuthorship W3203374547A5027438317 @default.
- W3203374547 hasAuthorship W3203374547A5030878947 @default.
- W3203374547 hasAuthorship W3203374547A5061529715 @default.
- W3203374547 hasAuthorship W3203374547A5084983727 @default.
- W3203374547 hasAuthorship W3203374547A5085963057 @default.
- W3203374547 hasConcept C115961682 @default.
- W3203374547 hasConcept C119857082 @default.
- W3203374547 hasConcept C124101348 @default.
- W3203374547 hasConcept C134306372 @default.
- W3203374547 hasConcept C153180895 @default.
- W3203374547 hasConcept C154945302 @default.
- W3203374547 hasConcept C177264268 @default.
- W3203374547 hasConcept C199360897 @default.
- W3203374547 hasConcept C2776145971 @default.
- W3203374547 hasConcept C33923547 @default.
- W3203374547 hasConcept C36503486 @default.
- W3203374547 hasConcept C41008148 @default.
- W3203374547 hasConcept C51632099 @default.
- W3203374547 hasConcept C58489278 @default.
- W3203374547 hasConcept C67186912 @default.
- W3203374547 hasConcept C739882 @default.
- W3203374547 hasConcept C77088390 @default.
- W3203374547 hasConcept C79337645 @default.
- W3203374547 hasConcept C97931131 @default.
- W3203374547 hasConcept C99498987 @default.
- W3203374547 hasConceptScore W3203374547C115961682 @default.
- W3203374547 hasConceptScore W3203374547C119857082 @default.
- W3203374547 hasConceptScore W3203374547C124101348 @default.
- W3203374547 hasConceptScore W3203374547C134306372 @default.
- W3203374547 hasConceptScore W3203374547C153180895 @default.
- W3203374547 hasConceptScore W3203374547C154945302 @default.
- W3203374547 hasConceptScore W3203374547C177264268 @default.
- W3203374547 hasConceptScore W3203374547C199360897 @default.
- W3203374547 hasConceptScore W3203374547C2776145971 @default.
- W3203374547 hasConceptScore W3203374547C33923547 @default.
- W3203374547 hasConceptScore W3203374547C36503486 @default.
- W3203374547 hasConceptScore W3203374547C41008148 @default.
- W3203374547 hasConceptScore W3203374547C51632099 @default.
- W3203374547 hasConceptScore W3203374547C58489278 @default.
- W3203374547 hasConceptScore W3203374547C67186912 @default.
- W3203374547 hasConceptScore W3203374547C739882 @default.
- W3203374547 hasConceptScore W3203374547C77088390 @default.
- W3203374547 hasConceptScore W3203374547C79337645 @default.
- W3203374547 hasConceptScore W3203374547C97931131 @default.
- W3203374547 hasConceptScore W3203374547C99498987 @default.
- W3203374547 hasFunder F4320311039 @default.
- W3203374547 hasFunder F4320335777 @default.
- W3203374547 hasLocation W32033745471 @default.
- W3203374547 hasOpenAccess W3203374547 @default.
- W3203374547 hasPrimaryLocation W32033745471 @default.
- W3203374547 hasRelatedWork W2030388591 @default.
- W3203374547 hasRelatedWork W2090668960 @default.
- W3203374547 hasRelatedWork W2160314512 @default.
- W3203374547 hasRelatedWork W2172289703 @default.
- W3203374547 hasRelatedWork W2251047611 @default.
- W3203374547 hasRelatedWork W2359185137 @default.
- W3203374547 hasRelatedWork W3006470798 @default.
- W3203374547 hasRelatedWork W4225000479 @default.
- W3203374547 hasRelatedWork W4225502772 @default.
- W3203374547 hasRelatedWork W2186522517 @default.
- W3203374547 isParatext "false" @default.
- W3203374547 isRetracted "false" @default.