MP2D: An Automated Topic Shift Dialogue Generation Framework Leveraging Knowledge Graphs
This addresses the problem of limited training data for topic shift dialogues in conversational AI, though it appears incremental as it builds on existing passage-to-dialogue methods.
The paper tackles the challenge of generating dialogue datasets with natural topic shifts by proposing MP2D, a framework that uses knowledge graphs to map topic transitions and creates conversational datasets; it shows that models trained on MP2D-generated data improve performance on topic shift tasks, and even large language models struggle with these shifts.
Despite advancements in on-topic dialogue systems, effectively managing topic shifts within dialogues remains a persistent challenge, largely attributed to the limited availability of training datasets. To address this issue, we propose Multi-Passage to Dialogue (MP2D), a data generation framework that automatically creates conversational question-answering datasets with natural topic transitions. By leveraging the relationships between entities in a knowledge graph, MP2D maps the flow of topics within a dialogue, effectively mirroring the dynamics of human conversation. It retrieves relevant passages corresponding to the topics and transforms them into dialogues through the passage-to-dialogue method. Through quantitative and qualitative experiments, we demonstrate MP2D's efficacy in generating dialogue with natural topic shifts. Furthermore, this study introduces a novel benchmark for topic shift dialogues, TS-WikiDialog. Utilizing the dataset, we demonstrate that even Large Language Models (LLMs) struggle to handle topic shifts in dialogue effectively, and we showcase the performance improvements of models trained on datasets generated by MP2D across diverse topic shift dialogue tasks.