AirConcierge: Generating Task-Oriented Dialogue via Efficient Large-Scale Knowledge Retrieval

Findings of the Association for Computational Linguistics: EMNLP 2020, pages 884–897 November 16 - 20, 2020. c⃝2020 Association for Computational Linguistics 884 AirConcierge: Generating Task-Oriented Dialogue via Efficient Large-Scale Knowledge Retrieval Chieh-Yang Chen† Pei-Hsin Wang† Shih-Chieh Chang† Da-Cheng Juan¶ Wei Wei¶ Jia-Yu Pan¶ †National Tsing-Hua University ¶Google Research {darius107062542, peihsin}@gapp.nthu.edu.tw scchang@cs.nthu.edu.tw {dacheng, wewei, jypan}@google.com Abstract Despite recent success in neural task-oriented dialogue systems, developing such a real- world system involves accessing large-scale knowledge bases (KBs), which cannot be sim- ply encoded by neural approaches, such as memory network mechanisms. To alleviate the above problem, we propose AirConcierge, an end-to-end trainable text-to-SQL guided framework to learn a neural agent that interacts with KBs using the generated SQL queries. Specifically, the neural agent first learns to ask and confirm the customer’s intent during the multi-turn interactions, then dynamically de- termining when to ground the user constraints into executable SQL queries so as to fetch rel- evant information from KBs. With the help of our method, the agent can use less but more accurate fetched results to generate use- ful responses efficiently, instead of incorporat- ing the entire KBs. We evaluate the proposed method on the AirDialogue dataset, a large cor- pus released by Google, containing the con- versations of customers booking flight tickets from the agent. The experimental results show that AirConcierge significantly improves over previous work in terms of accuracy and the BLEU score, which demonstrates not only the ability to achieve the given task but also the good quality of the generated dialogues. 1 Introduction The task-oriented dialogue system (Young et al., 2013) is one of the rapidly growing fields with many practical applications, attracting more and more research attention recently (Zhao and Esk´enazi, 2016; Wen et al., 2016; Bordes et al., 2017; Dhingra et al., 2017; Eric and Manning, 2017; Liu and Lane, 2017). In order to assist users in solving a specific task while holding conversa- tions with human, the agent needs to understand the intentions of a user during the conversation and Figure 1: An example of the task-oriented dialogue that incorporates a knowledge base (KB) from the AirDia- logue dataset. The agent ground the user constraints into executable SQL query at the turn annotated in red. fulfills the request. Such a process often involves interacting with external KBs to access task-related information. Figure 1 shows an example of a task- oriented dialogue between a user and an airline ticket reservation agent. Traditional dialogue systems (Kim et al., 2008; Deoras and Sarikaya, 2013) may rely on the prede- fined slot-filling pairs, where a set of slots needs to be filled during the conversation. In addition, some works (Sukhbaatar et al., 2015; Madotto et al., 2018; Wu et al., 2019) have considered integrating KBs in a task-oriented dialogue system to gener- ate a suitable response and have achieved promis- ing performance. However, these methods either are limited by predefined configurations or do not scale to large KBs. Since real-world KBs typically contain millions of records, end-to-end dialogue systems …

Content truncated to 500 words