Unsupervised Dialogue State Tracking for End-to-End Task-Oriented Dialogue with a Multi-Span Prediction Network

作     者：刘庆斌 何世柱 刘操 刘康 赵军 Qing-Bin Liu;Shi-Zhu He;Cao Liu;Kang Liu;Jun Zhao

作者机构：National Laboratory of Pattern RecognitionInstitute of AutomationChinese Academy of SciencesBeijing 100190China School of Artificial IntelligenceUniversity of Chinese Academy of SciencesBeijing 100049China Beijing Sankuai Online Technology Company LimitedBeijing 100102China 

出 版 物：《Journal of Computer Science & Technology》 (计算机科学技术学报（英文版）)

年 卷 期：2023年第38卷第4期

页      面：834-852页

核心收录：

学科分类：12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 081104[工学-模式识别与智能系统] 08[工学] 0835[工学-软件工程] 0811[工学-控制科学与工程] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：supported by the National Key Research and Development Program of China under Grant No.2020AAA0106400 the National Natural Science Foundation of China under Grant Nos.61922085 and 61976211 the Independent Research Project of National Laboratory of Pattern Recognition under Grant No.Z-2018013 the Key Research Program of Chinese Academy of Sciences(CAS)under Grant No.ZDBS-SSW-JSC006 the Youth Innovation Promotion Association CAS under Grant No.201912 

主　　题：end-to-end task-oriented dialogue dialogue state tracking(DST) unsupervised learning reinforcement learning 

摘      要：This paper focuses on end-to-end task-oriented dialogue systems,which jointly handle dialogue state tracking(DST)and response *** methods usually adopt a supervised paradigm to learn DST from a manually labeled ***,the annotation of the corpus is costly,time-consuming,and cannot cover a wide range of domains in the real *** solve this problem,we propose a multi-span prediction network(MSPN)that performs unsupervised DST for end-to-end task-oriented ***,MSPN contains a novel split-merge copy mechanism that captures long-term dependencies in dialogues to automatically extract multiple text spans as *** on these keywords,MSPN uses a semantic distance based clustering approach to obtain the values of each *** addition,we propose an ontology-based reinforcement learning approach,which employs the values of each slot to train MSPN to generate relevant *** results on single-domain and multi-domain task-oriented dialogue datasets show that MSPN achieves state-of-the-art performance with significant ***,we construct a new Chinese dialogue dataset MeDial in the low-resource medical domain,which further demonstrates the adaptability of MSPN.