CoPrUS: Consistency Preserving Utterance Synthesis towards more realistic benchmark dialogues
This work addresses the problem of unrealistic dialogue benchmarks for researchers and developers in dialogue systems, though it is incremental as it builds on existing datasets and methods.
The authors tackled the lack of realistic communication errors in benchmark dialogue datasets by developing CoPrUS, an automatic pipeline using LLMs to synthesize miscommunications like misunderstandings and non-understandings, resulting in a modified dataset of nearly 1900 dialogues for data augmentation.
Large-scale Wizard-Of-Oz dialogue datasets have enabled the training of deep learning-based dialogue systems. While they are successful as benchmark datasets, they lack certain types of utterances, which would make them more realistic. In this work, we investigate the creation of synthetic communication errors in an automatic pipeline. Based on linguistic theory, we propose and follow a simple error taxonomy. We focus on three types of miscommunications that could happen in real-world dialogues but are underrepresented in the benchmark dataset: misunderstandings, non-understandings and vaguely related questions. Our two-step approach uses a state-of-the-art Large Language Model (LLM) to first create the error and secondly the repairing utterance. We perform Language Model-based evaluation to ensure the quality of the generated utterances. We apply the method to the MultiWOZ dataset and evaluate it both qualitatively and empirically as well as with human judges. Our results indicate that current LLMs can aid in adding post-hoc miscommunications to benchmark datasets as a form of data augmentation. We publish the resulting dataset, in which nearly 1900 dialogues have been modified, as CoPrUS-MultiWOZ to facilitate future work on dialogue systems.