Matches in SemOpenAlex for { <https://semopenalex.org/work/W2890940245> ?p ?o ?g. }
- W2890940245 abstract "Combining the virtues of probability graphic models and neural networks, Conditional Variational Auto-encoder (CVAE) has shown promising performance in applications such as response generation. However, existing CVAE-based models often generate responses from a single latent variable which may not be sufficient to model high variability in responses. To solve this problem, we propose a novel model that sequentially introduces a series of latent variables to condition the generation of each word in the response sequence. In addition, the approximate posteriors of these latent variables are augmented with a backward Recurrent Neural Network (RNN), which allows the latent variables to capture long-term dependencies of future tokens in generation. To facilitate training, we supplement our model with an auxiliary objective that predicts the subsequent bag of words. Empirical experiments conducted on Opensubtitle and Reddit datasets show that the proposed model leads to significant improvement on both relevance and diversity over state-of-the-art baselines." @default.
- W2890940245 created "2018-09-27" @default.
- W2890940245 creator A5007348590 @default.
- W2890940245 creator A5015709853 @default.
- W2890940245 creator A5026719663 @default.
- W2890940245 creator A5049098538 @default.
- W2890940245 creator A5056630426 @default.
- W2890940245 creator A5086674741 @default.
- W2890940245 date "2018-01-01" @default.
- W2890940245 modified "2023-10-16" @default.
- W2890940245 title "Variational Autoregressive Decoder for Neural Response Generation" @default.
- W2890940245 cites W1518951372 @default.
- W2890940245 cites W1533861849 @default.
- W2890940245 cites W1924770834 @default.
- W2890940245 cites W2108501770 @default.
- W2890940245 cites W2130942839 @default.
- W2890940245 cites W2328886022 @default.
- W2890940245 cites W2560512785 @default.
- W2890940245 cites W2587284713 @default.
- W2890940245 cites W2593696076 @default.
- W2890940245 cites W2604444020 @default.
- W2890940245 cites W2756487349 @default.
- W2890940245 cites W2962717182 @default.
- W2890940245 cites W2962796276 @default.
- W2890940245 cites W2962974924 @default.
- W2890940245 cites W2963035145 @default.
- W2890940245 cites W2963206148 @default.
- W2890940245 cites W2963223306 @default.
- W2890940245 cites W2963279312 @default.
- W2890940245 cites W2963790827 @default.
- W2890940245 cites W2963958388 @default.
- W2890940245 cites W2964076986 @default.
- W2890940245 cites W2964121744 @default.
- W2890940245 cites W2964178377 @default.
- W2890940245 cites W2964339599 @default.
- W2890940245 cites W2964352131 @default.
- W2890940245 cites W592244745 @default.
- W2890940245 doi "https://doi.org/10.18653/v1/d18-1354" @default.
- W2890940245 hasPublicationYear "2018" @default.
- W2890940245 type Work @default.
- W2890940245 sameAs 2890940245 @default.
- W2890940245 citedByCount "43" @default.
- W2890940245 countsByYear W28909402452018 @default.
- W2890940245 countsByYear W28909402452019 @default.
- W2890940245 countsByYear W28909402452020 @default.
- W2890940245 countsByYear W28909402452021 @default.
- W2890940245 countsByYear W28909402452023 @default.
- W2890940245 crossrefType "proceedings-article" @default.
- W2890940245 hasAuthorship W2890940245A5007348590 @default.
- W2890940245 hasAuthorship W2890940245A5015709853 @default.
- W2890940245 hasAuthorship W2890940245A5026719663 @default.
- W2890940245 hasAuthorship W2890940245A5049098538 @default.
- W2890940245 hasAuthorship W2890940245A5056630426 @default.
- W2890940245 hasAuthorship W2890940245A5086674741 @default.
- W2890940245 hasBestOaLocation W28909402451 @default.
- W2890940245 hasConcept C11413529 @default.
- W2890940245 hasConcept C119857082 @default.
- W2890940245 hasConcept C143724316 @default.
- W2890940245 hasConcept C147168706 @default.
- W2890940245 hasConcept C149782125 @default.
- W2890940245 hasConcept C151730666 @default.
- W2890940245 hasConcept C154945302 @default.
- W2890940245 hasConcept C159877910 @default.
- W2890940245 hasConcept C2524010 @default.
- W2890940245 hasConcept C2778112365 @default.
- W2890940245 hasConcept C33923547 @default.
- W2890940245 hasConcept C41008148 @default.
- W2890940245 hasConcept C50644808 @default.
- W2890940245 hasConcept C51167844 @default.
- W2890940245 hasConcept C54355233 @default.
- W2890940245 hasConcept C65965080 @default.
- W2890940245 hasConcept C86803240 @default.
- W2890940245 hasConcept C90805587 @default.
- W2890940245 hasConceptScore W2890940245C11413529 @default.
- W2890940245 hasConceptScore W2890940245C119857082 @default.
- W2890940245 hasConceptScore W2890940245C143724316 @default.
- W2890940245 hasConceptScore W2890940245C147168706 @default.
- W2890940245 hasConceptScore W2890940245C149782125 @default.
- W2890940245 hasConceptScore W2890940245C151730666 @default.
- W2890940245 hasConceptScore W2890940245C154945302 @default.
- W2890940245 hasConceptScore W2890940245C159877910 @default.
- W2890940245 hasConceptScore W2890940245C2524010 @default.
- W2890940245 hasConceptScore W2890940245C2778112365 @default.
- W2890940245 hasConceptScore W2890940245C33923547 @default.
- W2890940245 hasConceptScore W2890940245C41008148 @default.
- W2890940245 hasConceptScore W2890940245C50644808 @default.
- W2890940245 hasConceptScore W2890940245C51167844 @default.
- W2890940245 hasConceptScore W2890940245C54355233 @default.
- W2890940245 hasConceptScore W2890940245C65965080 @default.
- W2890940245 hasConceptScore W2890940245C86803240 @default.
- W2890940245 hasConceptScore W2890940245C90805587 @default.
- W2890940245 hasLocation W28909402451 @default.
- W2890940245 hasOpenAccess W2890940245 @default.
- W2890940245 hasPrimaryLocation W28909402451 @default.
- W2890940245 hasRelatedWork W2234122438 @default.
- W2890940245 hasRelatedWork W2772692325 @default.
- W2890940245 hasRelatedWork W2890940245 @default.
- W2890940245 hasRelatedWork W2950067852 @default.
- W2890940245 hasRelatedWork W2962930338 @default.
- W2890940245 hasRelatedWork W4281386417 @default.