GL-GIN: Fast and Accurate Non-Autoregressive Model for Joint Multiple Intent Detection and Slot Filling

Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing, pages 178–188 August 1–6, 2021. ©2021 Association for Computational Linguistics 178 GL-GIN: Fast and Accurate Non-Autoregressive Model for Joint Multiple Intent Detection and Slot Filling Libo Qin, Fuxuan Wei, Tianbao Xie, Xiao Xu, Wanxiang Che∗, Ting Liu Research Center for Social Computing and Information Retrieval Harbin Institute of Technology, China {lbqin,fuxuanwei,tianbaoxie,xxu,car,tliu}@ir.hit.edu.cn Abstract Multi-intent SLU can handle multiple intents in an utterance, which has attracted increas- ing attention. However, the state-of-the-art joint models heavily rely on autoregressive ap- proaches, resulting in two issues: slow infer- ence speed and information leakage. In this paper, we explore a non-autoregressive model for joint multiple intent detection and slot fill- ing, achieving more fast and accurate. Specif- ically, we propose a Global-Locally Graph Interaction Network (GL-GIN) where a local slot-aware graph interaction layer is proposed to model slot dependency for alleviating unco- ordinated slots problem while a global intent- slot graph interaction layer is introduced to model the interaction between multiple intents and all slots in the utterance. Experimen- tal results on two public datasets show that our framework achieves state-of-the-art perfor- mance while being 11.5 times faster. 1 Introduction Spoken Language Understanding (SLU) (Young et al., 2013) is a critical component in spoken dialog systems, which aims to understand user’s queries. It typically includes two sub-tasks: intent detection and slot filling (Tur and De Mori, 2011). Since intents and slots are closely tied, dominant single-intent SLU systems in the literature (Goo et al., 2018; Li et al., 2018; Liu et al., 2019b; E et al., 2019; Qin et al., 2019; Teng et al., 2021; Qin et al., 2021b,c) adopt joint models to consider the correlation between the two tasks, which have obtained remarkable success. Multi-intent SLU means that the system can han- dle an utterance containing multiple intents, which is shown to be more practical in the real-world sce- nario, attracting increasing attention. To this end, ∗Corresponding author. (a) How far is Denver airport (b) How far is Denver airport O O O B-AN I-AN O O O B-AN I-AN Figure 1: (a) Autoregressive model generates outputs word by word from left-to-right direction. The gray color denotes the unseen information when model de- codes for the word Denver. (b) Non-autoregressive model can produce outputs in parallel. AN denotes airport name. Xu and Sarikaya (2013) and Kim et al. (2017) be- gin to explore the multi-intent SLU. However, their models only consider the multiple intent detection while ignoring slot filling task. Recently, Gangad- haraiah and Narayanaswamy (2019) make the first attempt to propose a multi-task framework to joint model the multiple intent detection and slot filling. Qin et al. (2020b) further propose an adaptive inter- action framework (AGIF) to achieve fine-grained multi-intent information integration for slot filling, obtaining state-of-the-art performance. Though achieving the promising performance, the existing multi-intent SLU joint models heav- ily rely on an autoregressive fashion, as shown in Figure 1(a), …

Content truncated to 500 words