Matches in SemOpenAlex for { <https://semopenalex.org/work/W3202125623> ?p ?o ?g. }
- W3202125623 abstract "Learning sentence embeddings from dialogues has drawn increasing attention due to its low annotation cost and high domain adaptability. Conventional approaches employ the siamese-network for this task, which obtains the sentence embeddings through modeling the context-response semantic relevance by applying a feed-forward network on top of the sentence encoders. However, as the semantic textual similarity is commonly measured through the element-wise distance metrics (e.g. cosine and L2 distance), such architecture yields a large gap between training and evaluating. In this paper, we propose DialogueCSE, a dialogue-based contrastive learning approach to tackle this issue. DialogueCSE first introduces a novel matching-guided embedding (MGE) mechanism, which generates a context-aware embedding for each candidate response embedding (i.e. the context-free embedding) according to the guidance of the multi-turn context-response matching matrices. Then it pairs each context-aware embedding with its corresponding context-free embedding and finally minimizes the contrastive loss across all pairs. We evaluate our model on three multi-turn dialogue datasets: the Microsoft Dialogue Corpus, the Jing Dong Dialogue Corpus, and the E-commerce Dialogue Corpus. Evaluation results show that our approach significantly outperforms the baselines across all three datasets in terms of MAP and Spearman’s correlation measures, demonstrating its effectiveness. Further quantitative experiments show that our approach achieves better performance when leveraging more dialogue context and remains robust when less training data is provided." @default.
- W3202125623 created "2021-10-11" @default.
- W3202125623 creator A5013292507 @default.
- W3202125623 creator A5014579261 @default.
- W3202125623 creator A5038647128 @default.
- W3202125623 creator A5046872618 @default.
- W3202125623 creator A5055526924 @default.
- W3202125623 creator A5070812231 @default.
- W3202125623 date "2021-01-01" @default.
- W3202125623 modified "2023-09-27" @default.
- W3202125623 title "DialogueCSE: Dialogue-based Contrastive Learning of Sentence Embeddings" @default.
- W3202125623 cites W1486649854 @default.
- W3202125623 cites W1665214252 @default.
- W3202125623 cites W1840435438 @default.
- W3202125623 cites W2155482025 @default.
- W3202125623 cites W2250539671 @default.
- W3202125623 cites W2250790822 @default.
- W3202125623 cites W2413533759 @default.
- W3202125623 cites W2611029872 @default.
- W3202125623 cites W2790235966 @default.
- W3202125623 cites W2798583685 @default.
- W3202125623 cites W2842511635 @default.
- W3202125623 cites W2891177506 @default.
- W3202125623 cites W2949433733 @default.
- W3202125623 cites W2953384591 @default.
- W3202125623 cites W2963310665 @default.
- W3202125623 cites W2963341956 @default.
- W3202125623 cites W2963644595 @default.
- W3202125623 cites W2963804993 @default.
- W3202125623 cites W2963846996 @default.
- W3202125623 cites W2963918774 @default.
- W3202125623 cites W3031414376 @default.
- W3202125623 cites W3033406728 @default.
- W3202125623 cites W3034238904 @default.
- W3202125623 cites W3100652389 @default.
- W3202125623 cites W3104033643 @default.
- W3202125623 cites W3104078590 @default.
- W3202125623 cites W3105816068 @default.
- W3202125623 cites W3115295967 @default.
- W3202125623 cites W3131870090 @default.
- W3202125623 cites W3154229486 @default.
- W3202125623 cites W3156636935 @default.
- W3202125623 cites W3164054899 @default.
- W3202125623 cites W3193735869 @default.
- W3202125623 doi "https://doi.org/10.18653/v1/2021.emnlp-main.185" @default.
- W3202125623 hasPublicationYear "2021" @default.
- W3202125623 type Work @default.
- W3202125623 sameAs 3202125623 @default.
- W3202125623 citedByCount "3" @default.
- W3202125623 countsByYear W32021256232022 @default.
- W3202125623 crossrefType "proceedings-article" @default.
- W3202125623 hasAuthorship W3202125623A5013292507 @default.
- W3202125623 hasAuthorship W3202125623A5014579261 @default.
- W3202125623 hasAuthorship W3202125623A5038647128 @default.
- W3202125623 hasAuthorship W3202125623A5046872618 @default.
- W3202125623 hasAuthorship W3202125623A5055526924 @default.
- W3202125623 hasAuthorship W3202125623A5070812231 @default.
- W3202125623 hasBestOaLocation W32021256231 @default.
- W3202125623 hasConcept C103278499 @default.
- W3202125623 hasConcept C105795698 @default.
- W3202125623 hasConcept C115961682 @default.
- W3202125623 hasConcept C151730666 @default.
- W3202125623 hasConcept C153180895 @default.
- W3202125623 hasConcept C154945302 @default.
- W3202125623 hasConcept C158154518 @default.
- W3202125623 hasConcept C165064840 @default.
- W3202125623 hasConcept C17744445 @default.
- W3202125623 hasConcept C199539241 @default.
- W3202125623 hasConcept C204321447 @default.
- W3202125623 hasConcept C2777530160 @default.
- W3202125623 hasConcept C2779343474 @default.
- W3202125623 hasConcept C2780762811 @default.
- W3202125623 hasConcept C33923547 @default.
- W3202125623 hasConcept C41008148 @default.
- W3202125623 hasConcept C41608201 @default.
- W3202125623 hasConcept C86803240 @default.
- W3202125623 hasConceptScore W3202125623C103278499 @default.
- W3202125623 hasConceptScore W3202125623C105795698 @default.
- W3202125623 hasConceptScore W3202125623C115961682 @default.
- W3202125623 hasConceptScore W3202125623C151730666 @default.
- W3202125623 hasConceptScore W3202125623C153180895 @default.
- W3202125623 hasConceptScore W3202125623C154945302 @default.
- W3202125623 hasConceptScore W3202125623C158154518 @default.
- W3202125623 hasConceptScore W3202125623C165064840 @default.
- W3202125623 hasConceptScore W3202125623C17744445 @default.
- W3202125623 hasConceptScore W3202125623C199539241 @default.
- W3202125623 hasConceptScore W3202125623C204321447 @default.
- W3202125623 hasConceptScore W3202125623C2777530160 @default.
- W3202125623 hasConceptScore W3202125623C2779343474 @default.
- W3202125623 hasConceptScore W3202125623C2780762811 @default.
- W3202125623 hasConceptScore W3202125623C33923547 @default.
- W3202125623 hasConceptScore W3202125623C41008148 @default.
- W3202125623 hasConceptScore W3202125623C41608201 @default.
- W3202125623 hasConceptScore W3202125623C86803240 @default.
- W3202125623 hasLocation W32021256231 @default.
- W3202125623 hasLocation W32021256232 @default.
- W3202125623 hasOpenAccess W3202125623 @default.
- W3202125623 hasPrimaryLocation W32021256231 @default.
- W3202125623 hasRelatedWork W159132833 @default.
- W3202125623 hasRelatedWork W19743771 @default.