To handle these phenomena, we propose a Dialogue State Tracking with Slot Connections (DST-SC) model to explicitly consider slot correlations throughout different domains. Specially, we first apply a Slot Attention to learn a set of slot-specific options from the original dialogue and then combine them using a slot data sharing module. Slot Attention with Value Normalization for Multi-Domain Dialogue State Tracking Yexiang Wang writer Yi Guo author Siqi Zhu author 2020-nov text Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP) Association for Computational Linguistics Online conference publication Incompleteness of area ontology and unavailability of some values are two inevitable problems of dialogue state tracking (DST). On this paper, we suggest a new structure to cleverly exploit ontology, which consists of Slot Attention (SA) and Value Normalization (VN), known as SAVN. SAS: Dialogue State Tracking through Slot Attention and Slot Information Sharing Jiaying Hu author Yan Yang writer Chencai Chen writer Liang He author Zhou Yu author 2020-jul textual content Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics Association for Computational Linguistics Online conference publication Dialogue state tracker is answerable for inferring user intentions through dialogue historical past. We propose a Dialogue State Tracker with Slot Attention and Slot Information Sharing (SAS) to scale back redundant information’s interference and improve long dialogue context monitoring.
