Leveraging Slot Descriptions for Zero-Shot Cross-Domain Dialogue StateTracking

Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 5640–5648 June 6–11, 2021. ©2021 Association for Computational Linguistics 5640 Leveraging Slot Descriptions for Zero-Shot Cross-Domain Dialogue State Tracking Zhaojiang Lin1∗, Bing Liu2, Seungwhan Moon2, Paul Crook2, Zhenpeng Zhou2, Zhiguang Wang2, Zhou Yu3, Andrea Madotto1∗, Eunjoon Cho2, Rajen Subba2 1The Hong Kong University of Science and Technology 2Facebook 3Columbia University zlinao@ust.hk, bingl@fb.com Abstract Zero-shot cross-domain dialogue state track- ing (DST) enables us to handle task-oriented dialogue in unseen domains without the ex- pense of collecting in-domain data. In this paper, we propose a slot description enhanced generative approach for zero-shot cross-domain DST. Specifically, our model first encodes dialogue context and slots with a pre-trained self-attentive encoder, and gener- ates slot values in an auto-regressive manner. In addition, we incorporate Slot Type Informed Descriptions that capture the shared informa- tion across slots to facilitate cross-domain knowledge transfer. Experimental results on the MultiWOZ dataset show that our proposed method significantly improves existing state- of-the-art results in the zero-shot cross-domain setting. 1 Introduction Task-oriented dialogue systems are designed to assist users in performing daily activities, such as restaurant booking, travel planning, and on- line shopping. These virtual assistants provide natural language interfaces to services and on- line APIs (Rastogi et al., 2020). Based on users’ needs, these systems frequently require support for new domains. However, the current state-of- the-art systems require a substantial amount of in- domain data to properly model a new domain. The data-collection process is both expensive and time- consuming, and thus it is very important to study methods that can build robust and scalable dialogue systems using little to no in-domain data. The dialogue state tracking (DST) is an essential component of task-oriented dialogue systems that tracks users’ requirements over multi-turn conver- sations. A popular formulation of the dialogue state is in the form of a list of slot-value pairs. In DST, tracking unseen slots in a new domain, a.k.a. zero- shot domain adaptation, is a significant challenge, ∗ Work done during internship at Facebook Figure 1: High-level description of the T5DST. The model takes dialogue history and slot name as input, and generates the value. since the model has never seen in-domain training samples. There are two main lines of work to tackle this problem. The first proposes domain trans- ferable models using copy mechanisms or ontol- ogy graph information (Wu et al., 2019; Zhou and Small, 2019). A limitation of such models is that they may not fully leverage pre-trained language models due to the specialized model architecture. The second line of work uses slot-descriptions as input to the model to facilitate the slot understand- ing (Rastogi et al., 2020). However, the provided slot descriptions are collected by crowd sourced hu- man annotators and might be inconsistent among different domains. In general, the optimal approach for constructing slot descriptions in zero-shot set- tings remains unexplored. In this work, we tackle the challenge of zero- shot cross-domain DST …

Content truncated to 500 words