Matches in SemOpenAlex for { <https://semopenalex.org/work/W3168264889> ?p ?o ?g. }
Showing items 1 to 90 of
90
with 100 items per page.
- W3168264889 abstract "Neural dialogue generation models trained with the one-hot target distribution suffer from the over-confidence issue, which leads to poor generation diversity as widely reported in the literature. Although existing approaches such as label smoothing can alleviate this issue, they fail to adapt to diverse dialog contexts. In this paper, we propose an Adaptive Label Smoothing (AdaLabel) approach that can adaptively estimate a target label distribution at each time step for different contexts. The maximum probability in the predicted distribution is used to modify the soft target distribution produced by a novel light-weight bi-directional decoder module. The resulting target distribution is aware of both previous and future contexts and is adjusted to avoid over-training the dialogue model. Our model can be trained in an end-to-end manner. Extensive experiments on two benchmark datasets show that our approach outperforms various competitive baselines in producing diverse responses." @default.
- W3168264889 created "2021-06-22" @default.
- W3168264889 creator A5033421780 @default.
- W3168264889 creator A5035186499 @default.
- W3168264889 creator A5044042138 @default.
- W3168264889 creator A5054920282 @default.
- W3168264889 date "2021-05-30" @default.
- W3168264889 modified "2023-10-12" @default.
- W3168264889 title "Diversifying Dialog Generation via Adaptive Label Smoothing" @default.
- W3168264889 cites W1821462560 @default.
- W3168264889 cites W1975879668 @default.
- W3168264889 cites W2101105183 @default.
- W3168264889 cites W2183341477 @default.
- W3168264889 cites W2626967530 @default.
- W3168264889 cites W2743473392 @default.
- W3168264889 cites W2788932366 @default.
- W3168264889 cites W2890940245 @default.
- W3168264889 cites W2916898195 @default.
- W3168264889 cites W2948210185 @default.
- W3168264889 cites W2951883832 @default.
- W3168264889 cites W2962717182 @default.
- W3168264889 cites W2963212250 @default.
- W3168264889 cites W2963341956 @default.
- W3168264889 cites W2964121744 @default.
- W3168264889 cites W2995404354 @default.
- W3168264889 cites W2995460523 @default.
- W3168264889 cites W2995969307 @default.
- W3168264889 cites W2997892440 @default.
- W3168264889 cites W3007260230 @default.
- W3168264889 cites W3015322406 @default.
- W3168264889 cites W3035239386 @default.
- W3168264889 cites W3087749337 @default.
- W3168264889 cites W3093956460 @default.
- W3168264889 cites W3100737666 @default.
- W3168264889 cites W3103182178 @default.
- W3168264889 cites W3125507956 @default.
- W3168264889 doi "https://doi.org/10.48550/arxiv.2105.14556" @default.
- W3168264889 hasPublicationYear "2021" @default.
- W3168264889 type Work @default.
- W3168264889 sameAs 3168264889 @default.
- W3168264889 citedByCount "0" @default.
- W3168264889 crossrefType "posted-content" @default.
- W3168264889 hasAuthorship W3168264889A5033421780 @default.
- W3168264889 hasAuthorship W3168264889A5035186499 @default.
- W3168264889 hasAuthorship W3168264889A5044042138 @default.
- W3168264889 hasAuthorship W3168264889A5054920282 @default.
- W3168264889 hasBestOaLocation W31682648891 @default.
- W3168264889 hasConcept C110121322 @default.
- W3168264889 hasConcept C119857082 @default.
- W3168264889 hasConcept C13280743 @default.
- W3168264889 hasConcept C134306372 @default.
- W3168264889 hasConcept C136764020 @default.
- W3168264889 hasConcept C154945302 @default.
- W3168264889 hasConcept C173853756 @default.
- W3168264889 hasConcept C185798385 @default.
- W3168264889 hasConcept C205649164 @default.
- W3168264889 hasConcept C31972630 @default.
- W3168264889 hasConcept C33923547 @default.
- W3168264889 hasConcept C3770464 @default.
- W3168264889 hasConcept C41008148 @default.
- W3168264889 hasConceptScore W3168264889C110121322 @default.
- W3168264889 hasConceptScore W3168264889C119857082 @default.
- W3168264889 hasConceptScore W3168264889C13280743 @default.
- W3168264889 hasConceptScore W3168264889C134306372 @default.
- W3168264889 hasConceptScore W3168264889C136764020 @default.
- W3168264889 hasConceptScore W3168264889C154945302 @default.
- W3168264889 hasConceptScore W3168264889C173853756 @default.
- W3168264889 hasConceptScore W3168264889C185798385 @default.
- W3168264889 hasConceptScore W3168264889C205649164 @default.
- W3168264889 hasConceptScore W3168264889C31972630 @default.
- W3168264889 hasConceptScore W3168264889C33923547 @default.
- W3168264889 hasConceptScore W3168264889C3770464 @default.
- W3168264889 hasConceptScore W3168264889C41008148 @default.
- W3168264889 hasLocation W31682648891 @default.
- W3168264889 hasOpenAccess W3168264889 @default.
- W3168264889 hasPrimaryLocation W31682648891 @default.
- W3168264889 hasRelatedWork W1485630101 @default.
- W3168264889 hasRelatedWork W1607341183 @default.
- W3168264889 hasRelatedWork W1862650538 @default.
- W3168264889 hasRelatedWork W2352008582 @default.
- W3168264889 hasRelatedWork W2368015273 @default.
- W3168264889 hasRelatedWork W2370437920 @default.
- W3168264889 hasRelatedWork W2498017833 @default.
- W3168264889 hasRelatedWork W2997818504 @default.
- W3168264889 hasRelatedWork W4200446208 @default.
- W3168264889 hasRelatedWork W628946606 @default.
- W3168264889 isParatext "false" @default.
- W3168264889 isRetracted "false" @default.
- W3168264889 magId "3168264889" @default.
- W3168264889 workType "article" @default.