Matches in SemOpenAlex for { <https://semopenalex.org/work/W4284683546> ?p ?o ?g. }
- W4284683546 abstract "Recently, pre-training methods have shown remarkable success in task-oriented dialog (TOD) systems. However, most existing pre-trained models for TOD focus on either dialog understanding or dialog generation, but not both. In this paper, we propose SPACE, a novel unified pre-trained dialog model learning from large-scale dialog corpora with limited annotations, which can be effectively fine-tuned on a wide range of downstream dialog tasks. Specifically, SPACE consists of four successive components in a single transformer to maintain a task-flow in TOD systems: (i) a dialog encoding module to encode dialog history, (ii) a dialog understanding module to extract semantic vectors from either user queries or system responses, (iii) a dialog policy module to generate a policy vector that contains high-level semantics of the response, and (iv) a dialog generation module to produce appropriate responses. We design a dedicated pre-training objective for each component. Concretely, we pre-train the dialog encoding module with span mask language modeling to learn contextualized dialog information. To capture the structured dialog semantics, we pre-train the dialog understanding module via a novel tree-induced semi-supervised contrastive learning objective with the help of extra dialog annotations. In addition, we pre-train the dialog policy module by minimizing the ℒ2 distance between its output policy vector and the semantic vector of the response for policy optimization. Finally, the dialog generation model is pre-trained by language modeling. Results show that SPACE achieves state-of-the-art performance on eight downstream dialog benchmarks, including intent prediction, dialog state tracking, and end-to-end dialog modeling. We also show that SPACE has a stronger few-shot ability than existing models under the low-resource setting." @default.
- W4284683546 created "2022-07-08" @default.
- W4284683546 creator A5013292507 @default.
- W4284683546 creator A5022173400 @default.
- W4284683546 creator A5030125216 @default.
- W4284683546 creator A5046872618 @default.
- W4284683546 creator A5054067088 @default.
- W4284683546 creator A5055526924 @default.
- W4284683546 creator A5086383656 @default.
- W4284683546 date "2022-07-06" @default.
- W4284683546 modified "2023-10-14" @default.
- W4284683546 title "Unified Dialog Model Pre-training for Task-Oriented Dialog Understanding and Generation" @default.
- W4284683546 cites W2128970689 @default.
- W4284683546 cites W2468710617 @default.
- W4284683546 cites W2786472750 @default.
- W4284683546 cites W2798914047 @default.
- W4284683546 cites W2889448364 @default.
- W4284683546 cites W2891704263 @default.
- W4284683546 cites W2891826200 @default.
- W4284683546 cites W2892248135 @default.
- W4284683546 cites W2898856000 @default.
- W4284683546 cites W2923014074 @default.
- W4284683546 cites W2952607215 @default.
- W4284683546 cites W2963748441 @default.
- W4284683546 cites W2963797754 @default.
- W4284683546 cites W2963825865 @default.
- W4284683546 cites W2964006684 @default.
- W4284683546 cites W2964223283 @default.
- W4284683546 cites W2970442801 @default.
- W4284683546 cites W2972156721 @default.
- W4284683546 cites W2972664115 @default.
- W4284683546 cites W2979520138 @default.
- W4284683546 cites W2988937804 @default.
- W4284683546 cites W2997771882 @default.
- W4284683546 cites W3011411500 @default.
- W4284683546 cites W3015217416 @default.
- W4284683546 cites W3034284249 @default.
- W4284683546 cites W3034861927 @default.
- W4284683546 cites W3034879520 @default.
- W4284683546 cites W3035451444 @default.
- W4284683546 cites W3045492832 @default.
- W4284683546 cites W3045703328 @default.
- W4284683546 cites W3098807778 @default.
- W4284683546 cites W3099590177 @default.
- W4284683546 cites W3100110884 @default.
- W4284683546 cites W3101223450 @default.
- W4284683546 cites W3102445752 @default.
- W4284683546 cites W3109620645 @default.
- W4284683546 cites W3112388682 @default.
- W4284683546 cites W3114038149 @default.
- W4284683546 cites W3117925982 @default.
- W4284683546 cites W3154271556 @default.
- W4284683546 cites W3155584966 @default.
- W4284683546 cites W3161101519 @default.
- W4284683546 cites W3176740717 @default.
- W4284683546 cites W4205633461 @default.
- W4284683546 cites W4205991051 @default.
- W4284683546 cites W4214567023 @default.
- W4284683546 doi "https://doi.org/10.1145/3477495.3532069" @default.
- W4284683546 hasPublicationYear "2022" @default.
- W4284683546 type Work @default.
- W4284683546 citedByCount "8" @default.
- W4284683546 countsByYear W42846835462022 @default.
- W4284683546 countsByYear W42846835462023 @default.
- W4284683546 crossrefType "proceedings-article" @default.
- W4284683546 hasAuthorship W4284683546A5013292507 @default.
- W4284683546 hasAuthorship W4284683546A5022173400 @default.
- W4284683546 hasAuthorship W4284683546A5030125216 @default.
- W4284683546 hasAuthorship W4284683546A5046872618 @default.
- W4284683546 hasAuthorship W4284683546A5054067088 @default.
- W4284683546 hasAuthorship W4284683546A5055526924 @default.
- W4284683546 hasAuthorship W4284683546A5086383656 @default.
- W4284683546 hasBestOaLocation W42846835461 @default.
- W4284683546 hasConcept C107457646 @default.
- W4284683546 hasConcept C136764020 @default.
- W4284683546 hasConcept C154945302 @default.
- W4284683546 hasConcept C162324750 @default.
- W4284683546 hasConcept C173853756 @default.
- W4284683546 hasConcept C184337299 @default.
- W4284683546 hasConcept C187736073 @default.
- W4284683546 hasConcept C190954187 @default.
- W4284683546 hasConcept C193125573 @default.
- W4284683546 hasConcept C199360897 @default.
- W4284683546 hasConcept C204321447 @default.
- W4284683546 hasConcept C2780451532 @default.
- W4284683546 hasConcept C41008148 @default.
- W4284683546 hasConceptScore W4284683546C107457646 @default.
- W4284683546 hasConceptScore W4284683546C136764020 @default.
- W4284683546 hasConceptScore W4284683546C154945302 @default.
- W4284683546 hasConceptScore W4284683546C162324750 @default.
- W4284683546 hasConceptScore W4284683546C173853756 @default.
- W4284683546 hasConceptScore W4284683546C184337299 @default.
- W4284683546 hasConceptScore W4284683546C187736073 @default.
- W4284683546 hasConceptScore W4284683546C190954187 @default.
- W4284683546 hasConceptScore W4284683546C193125573 @default.
- W4284683546 hasConceptScore W4284683546C199360897 @default.
- W4284683546 hasConceptScore W4284683546C204321447 @default.
- W4284683546 hasConceptScore W4284683546C2780451532 @default.
- W4284683546 hasConceptScore W4284683546C41008148 @default.
- W4284683546 hasFunder F4320321001 @default.