: NLP Vocabulary: dialogue state tracking

applied NLP > human-computer interaction > task-oriented dialogue > dialogue state tracking

dialogue state tracking

A task in task-oriented dialogue systems, which involves tracking of user’s goal and slot values based on dialogue context to accurately represent the current status or context of the conversation in order to facilitate more effective dialogue management and response generation. (Based on Maheshwari and Nangi, Pre-training and Data Augmentation for Dialogue State Tracking)

Dialogue State Tracking is a multi-class classification task which involves identifying the slot values for each (domain slot) pair at each dialogue turn based on a pre-defined ontology. (Zeng, He, Wang, Fu & Xu, 2024)

Dialogue state tracking (DST) is an essential component in a task-oriented dialogue system. (Qixiang, Dong, Mou, Wang, Zeng, Guo, Sun & Xu, 2022)
Dialogue State Tracking is regarded as a multiclass classification task based on a pre-defined ontology. (Zeng, He, Wang, Zeng, Wang, Xian & Xu, 2023)
Though Dialogue State Tracking (DST) is a core component of spoken dialogue systems recent work on this task mostly deals with chat corpora disregarding the discrepancies between spoken and written language. (Jacqmin, Druart, Estève, Favre, M Rojas & Vielzeuf, 2023)
We evaluate different specialization models across several target languages on the intrinsic word similarity task and three downstream language understanding tasks where distinguishing between true semantic similarity and conceptual relatedness is crucial: dialog state tracking lexical simplification and semantic textual similarity. (Ponti, Vulic, Glavaš, Reichart & Korhonen, 2019)

http://data.loterre.fr/ark:/67375/8LP-PN68FK78-1

RDF/XML TURTLE JSON-LD Last modified 5/21/24

Loterre