FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2590513900> ?p ?o ?g. }

• W2590513900 abstract "People speak at different levels of specificity in different situations. Depending on their knowledge, interlocutors, mood, etc.} A conversational agent should have this ability and know when to be specific and when to be general. We propose an approach that gives a neural network--based conversational agent this ability. Our approach involves alternating between emph{data distillation} and model training : removing training examples that are closest to the responses most commonly produced by the model trained from the last round and then retrain the model on the remaining dataset. Dialogue generation models trained with different degrees of data distillation manifest different levels of specificity. We then train a reinforcement learning system for selecting among this pool of generation models, to choose the best level of specificity for a given input. Compared to the original generative model trained without distillation, the proposed system is capable of generating more interesting and higher-quality responses, in addition to appropriately adjusting specificity depending on the context. Our research constitutes a specific case of a broader approach involving training multiple subsystems from a single dataset distinguished by differences in a specific property one wishes to model. We show that from such a set of subsystems, one can use reinforcement learning to build a system that tailors its output to different input contexts at test time." @default.
• W2590513900 created "2017-03-03" @default.
• W2590513900 creator A5001666093 @default.
• W2590513900 creator A5077851706 @default.
• W2590513900 creator A5087088138 @default.
• W2590513900 date "2017-02-22" @default.
• W2590513900 modified "2023-09-25" @default.
• W2590513900 title "Data Distillation for Controlling Specificity in Dialogue Generation" @default.
• W2590513900 cites W10957333 @default.
• W2590513900 cites W1518951372 @default.
• W2590513900 cites W1591706642 @default.
• W2590513900 cites W1902237438 @default.
• W2590513900 cites W1912128066 @default.
• W2590513900 cites W2008074204 @default.
• W2590513900 cites W2140641199 @default.
• W2590513900 cites W2146861498 @default.
• W2590513900 cites W2176263492 @default.
• W2590513900 cites W2204302769 @default.
• W2590513900 cites W2210838531 @default.
• W2590513900 cites W2328886022 @default.
• W2590513900 cites W2399060250 @default.
• W2590513900 cites W2410983263 @default.
• W2590513900 cites W2550893117 @default.
• W2590513900 cites W2551884415 @default.
• W2590513900 cites W2562579542 @default.
• W2590513900 cites W2583679610 @default.
• W2590513900 cites W2627074894 @default.
• W2590513900 cites W2734812719 @default.
• W2590513900 cites W2950726992 @default.
• W2590513900 cites W2952001554 @default.
• W2590513900 cites W2953135395 @default.
• W2590513900 cites W2962883855 @default.
• W2590513900 cites W2963206148 @default.
• W2590513900 cites W2963790827 @default.
• W2590513900 cites W2963986868 @default.
• W2590513900 cites W2964308564 @default.
• W2590513900 doi "https://doi.org/10.48550/arxiv.1702.06703" @default.
• W2590513900 hasPublicationYear "2017" @default.
• W2590513900 type Work @default.
• W2590513900 sameAs 2590513900 @default.
• W2590513900 citedByCount "13" @default.
• W2590513900 countsByYear W25905139002018 @default.
• W2590513900 countsByYear W25905139002019 @default.
• W2590513900 countsByYear W25905139002020 @default.
• W2590513900 countsByYear W25905139002021 @default.
• W2590513900 crossrefType "posted-content" @default.
• W2590513900 hasAuthorship W2590513900A5001666093 @default.
• W2590513900 hasAuthorship W2590513900A5077851706 @default.
• W2590513900 hasAuthorship W2590513900A5087088138 @default.
• W2590513900 hasBestOaLocation W25905139001 @default.
• W2590513900 hasConcept C111472728 @default.
• W2590513900 hasConcept C119857082 @default.
• W2590513900 hasConcept C138885662 @default.
• W2590513900 hasConcept C151730666 @default.
• W2590513900 hasConcept C154945302 @default.
• W2590513900 hasConcept C167966045 @default.
• W2590513900 hasConcept C177264268 @default.
• W2590513900 hasConcept C178790620 @default.
• W2590513900 hasConcept C185592680 @default.
• W2590513900 hasConcept C189950617 @default.
• W2590513900 hasConcept C199360897 @default.
• W2590513900 hasConcept C204030448 @default.
• W2590513900 hasConcept C2779343474 @default.
• W2590513900 hasConcept C39890363 @default.
• W2590513900 hasConcept C41008148 @default.
• W2590513900 hasConcept C50644808 @default.
• W2590513900 hasConcept C51632099 @default.
• W2590513900 hasConcept C86803240 @default.
• W2590513900 hasConcept C97541855 @default.
• W2590513900 hasConceptScore W2590513900C111472728 @default.
• W2590513900 hasConceptScore W2590513900C119857082 @default.
• W2590513900 hasConceptScore W2590513900C138885662 @default.
• W2590513900 hasConceptScore W2590513900C151730666 @default.
• W2590513900 hasConceptScore W2590513900C154945302 @default.
• W2590513900 hasConceptScore W2590513900C167966045 @default.
• W2590513900 hasConceptScore W2590513900C177264268 @default.
• W2590513900 hasConceptScore W2590513900C178790620 @default.
• W2590513900 hasConceptScore W2590513900C185592680 @default.
• W2590513900 hasConceptScore W2590513900C189950617 @default.
• W2590513900 hasConceptScore W2590513900C199360897 @default.
• W2590513900 hasConceptScore W2590513900C204030448 @default.
• W2590513900 hasConceptScore W2590513900C2779343474 @default.
• W2590513900 hasConceptScore W2590513900C39890363 @default.
• W2590513900 hasConceptScore W2590513900C41008148 @default.
• W2590513900 hasConceptScore W2590513900C50644808 @default.
• W2590513900 hasConceptScore W2590513900C51632099 @default.
• W2590513900 hasConceptScore W2590513900C86803240 @default.
• W2590513900 hasConceptScore W2590513900C97541855 @default.
• W2590513900 hasLocation W25905139001 @default.
• W2590513900 hasOpenAccess W2590513900 @default.
• W2590513900 hasPrimaryLocation W25905139001 @default.
• W2590513900 hasRelatedWork W2090093112 @default.
• W2590513900 hasRelatedWork W2890766902 @default.
• W2590513900 hasRelatedWork W2981889162 @default.
• W2590513900 hasRelatedWork W2987090991 @default.
• W2590513900 hasRelatedWork W3022038857 @default.
• W2590513900 hasRelatedWork W3129477419 @default.
• W2590513900 hasRelatedWork W3212218375 @default.
• W2590513900 hasRelatedWork W4287818197 @default.
• W2590513900 hasRelatedWork W4288026814 @default.

• next