Don't be Contradicted with Anything! CI-ToD: Towards Benchmarking Consistency for Task-oriented Dialogue System

TL;DR

This paper introduces CI-ToD, a new dataset for assessing consistency in task-oriented dialogue systems, highlighting the challenge of improving consistency identification beyond current state-of-the-art methods.

Contribution

The paper presents a novel dataset with fine-grained labels for consistency in task-oriented dialogue, and provides comprehensive experiments and analysis to guide future research.

Findings

State-of-the-art models achieve only 51.3% accuracy

Human performance on consistency identification is 93.2%

Room for improvement in consistency detection methods

Abstract

Consistency Identification has obtained remarkable success on open-domain dialogue, which can be used for preventing inconsistent response generation. However, in contrast to the rapid development in open-domain dialogue, few efforts have been made to the task-oriented dialogue direction. In this paper, we argue that consistency problem is more urgent in task-oriented domain. To facilitate the research, we introduce CI-ToD, a novel dataset for Consistency Identification in Task-oriented Dialog system. In addition, we not only annotate the single label to enable the model to judge whether the system response is contradictory, but also provide more fine-grained labels (i.e., Dialogue History Inconsistency, User Query Inconsistency and Knowledge Base Inconsistency) to encourage model to know what inconsistent sources lead to it. Empirical results show that state-of-the-art methods only achieve 51.3%, which is far behind the human performance of 93.2%, indicating that there is ample room for improving consistency identification ability. Finally, we conduct exhaustive experiments and qualitative analysis to comprehend key challenges and provide guidance for future directions. All datasets and models are publicly available at \url{https://github.com/yizhen20133868/CI-ToD}.