HyKnow: End-to-End Task-Oriented Dialog Modeling with Hybrid Knowledge Management

Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021, pages 1591–1602 August 1–6, 2021. ©2021 Association for Computational Linguistics 1591 HyKnow: End-to-End Task-Oriented Dialog Modeling with Hybrid Knowledge Management Silin Gao1∗, Ryuichi Takanobu1∗, Wei Peng2, Qun Liu2, Minlie Huang1† 1 CoAI Group, DCST, IAI, BNRIST, Tsinghua University, Beijing, China 2 Huawei Technologies, Shenzhen, China 1 gsl16@tsinghua.org.cn, gxly19@mails.tsinghua.edu.cn, aihuang@tsinghua.edu.cn 2 peng.wei1@huawei.com, qun.liu@huawei.com Abstract Task-oriented dialog (TOD) systems typically manage structured knowledge (e.g. ontologies and databases) to guide the goal-oriented con- versations. However, they fall short of han- dling dialog turns grounded on unstructured knowledge (e.g. reviews and documents). In this paper, we formulate a task of modeling TOD grounded on both structured and unstruc- tured knowledge. To address this task, we pro- pose a TOD system with hybrid knowledge management, HyKnow. It extends the belief state to manage both structured and unstruc- tured knowledge, and is the first end-to-end model that jointly optimizes dialog modeling grounded on these two kinds of knowledge. We conduct experiments on the modified ver- sion of MultiWOZ 2.1 dataset, where dialogs are grounded on hybrid knowledge. Experi- mental results show that HyKnow has strong end-to-end performance compared to existing TOD systems. It also outperforms the pipeline knowledge management schemes, with higher unstructured knowledge retrieval accuracy. 1 Introduction Recently, Task-Oriented Dialog (TOD) systems (Mehri et al., 2019; Zhang et al., 2020a,b; Le et al., 2020; Hosseini-Asl et al., 2020; Peng et al., 2020; Li et al., 2021) have achieved promising perfor- mance on accomplishing user goals. Most systems typically query structured knowledge such as ta- bles and databases based on the user goals, and use the query results to guide the generation of system responses, as shown in the first dialog turn in Fig. 1. However, real-world task-oriented conversations often step into dialog turns which are grounded on unstructured knowledge (Feng et al., 2020), such as passages and documents. For example, as the *Equal contribution. †Corresponding author. Figure 1: Illustration of task-oriented dialog modeling with hybrid knowledge management. Words in red and blue illustrate the new domain-slot-value triple and the topic of user utterance that we introduce into the belief state, respectively. Words in yellow illustrate the topics of documents that we extract through preprocessing. second dialog turn in Fig. 1 shows, the user asks about customers’ favorite food at Pizza Hut, which is grounded on the customer reviews of this restau- rant. Current TOD systems fall short of handling such dialog turns since they cannot utilize relevant unstructured knowledge. This deficiency may inter- rupt the dialog process, causing difficulties in track- ing user goals and generating system responses. In this work, we consider incorporating more various forms of domain knowledge into the TOD systems. Therefore, we define a task of model- ing TOD whose turns involve either structured or unstructured knowledge. In turns involving struc- 1592 tured knowledge, the system needs to track the user goals as triples and use them to perform database queries, whose results are used to generate the sys- tem response. While …

Content truncated to 500 words