Slot Attention with Value Normalization for Multi-Domain Dialogue State Tracking

Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing, pages 3019–3028, November 16–20, 2020. c⃝2020 Association for Computational Linguistics 3019 Slot Attention with Value Normalization for Multi-Domain Dialogue State Tracking Yexiang Wang, Yi Guo∗, and Siqi Zhu East China University of Science and Technology wyxiang@mail.ecust.edu.cn, guoyi@ecust.edu.cn zhusiqi@mail.ecust.edu.cn Abstract Incompleteness of domain ontology and un- availability of some values are two inevitable problems of dialogue state tracking (DST). Ex- isting approaches generally fall into two ex- tremes: choosing models without ontology or embedding ontology in models leading to over-dependence. In this paper, we propose a new architecture to cleverly exploit ontol- ogy, which consists of Slot Attention (SA) and Value Normalization (VN), referred to as SAVN. Moreover, we supplement the anno- tation of supporting span for MultiWOZ 2.1, which is the shortest span in utterances to sup- port the labeled value. SA shares knowledge between slots and utterances and only needs a simple structure to predict the supporting span. VN is designed specifically for the use of on- tology, which can convert supporting spans to the values. Empirical results demonstrate that SAVN achieves the state-of-the-art joint accu- racy of 54.52% on MultiWOZ 2.0 and 54.86% on MultiWOZ 2.1. Besides, we evaluate VN with incomplete ontology. The results show that even if only 30% ontology is used, VN can also contribute to our model. 1 Introduction Dialogue state tracking (DST) is a core component in the pipeline-based task-oriented dialog systems. The goal of DST is to extract the dialogue states which are indicated by a set of (domain, slot, value) triples during conversation. The (domain, slot, value) triple represents that previous conversation involves the slot of the domain and the specific con- tent is the value. For example, as shown in Figure 1, (restaurant, price, expensive) triple means that user wants to reserve an expensive restaurant. A high- quality DST model plays a significant role in the *Corresponding Author. dialogue system, because the dialogue states deter- mine the next system action (Chen et al., 2017). Traditional DST approaches generally rely on ontology already defined, where all slots and their possible values are given. With a predefined ontol- ogy, DST is simplified to a classification problem. The goal is to choose the most appropriate value from ontology for the slot (Mrkˇsi´c et al., 2017; Zhong et al., 2018). However, in practical applica- tions, a complete ontology is almost impossible to be defined in advance. To overcome the drawback, span-based (Xu and Hu, 2018; Gao et al., 2019) and generation (Wu et al., 2019) approaches spring up. The second problem is that some values required by DST cannot be found in utterances due to the di- verse descriptions during a conversation. As shown in Figure 1, value expensive was expressed as high end in the first turn. The problem gives rise to the powerlessness of span-based approaches. Recently, Zhang et al. (2019) show a dual strategy that com- bines the advantages of both the picklist-based and span-based methods. They use ontology …

Content truncated to 500 words