Transferable Persona-Grounded Dialogues via Grounded Minimal Edits
This work addresses the problem of improving persona consistency in dialogue systems for AI applications, though it is incremental as it builds on existing editing frameworks.
The paper tackles the transferability challenge of grounded dialogue models by proposing a grounded minimal editing framework that edits existing responses to incorporate new concepts, specifically personas, and demonstrates that their GME model outperforms baselines and improves persona consistency on external datasets.
Grounded dialogue models generate responses that are grounded on certain concepts. Limited by the distribution of grounded dialogue data, models trained on such data face the transferability challenges in terms of the data distribution and the type of grounded concepts. To address the challenges, we propose the grounded minimal editing framework, which minimally edits existing responses to be grounded on the given concept. Focusing on personas, we propose Grounded Minimal Editor (GME), which learns to edit by disentangling and recombining persona-related and persona-agnostic parts of the response. To evaluate persona-grounded minimal editing, we present the PersonaMinEdit dataset, and experimental results show that GME outperforms competitive baselines by a large margin. To evaluate the transferability, we experiment on the test set of BlendedSkillTalk and show that GME can edit dialogue models' responses to largely improve their persona consistency while preserving the use of knowledge and empathy.