Plug-and-Play Policy Planner for Large Language Model Powered Dialogue Agents

TL;DR

This paper introduces PPDPP, a plug-and-play framework for enhancing large language model dialogue agents' proactive capabilities through supervised fine-tuning and reinforcement learning, enabling better generalization and transferability.

Contribution

The paper proposes a novel plug-and-play dialogue policy planner for LLMs, combining supervised and reinforcement learning to improve proactive dialogue performance across applications.

Findings

PPDPP outperforms existing methods in negotiation, emotional support, and tutoring dialogues.

The framework enables generalization to new cases and applications by substituting the plug-in.

Experimental results show significant improvements in proactive dialogue tasks.

Abstract

Proactive dialogues serve as a practical yet challenging dialogue problem in the era of large language models (LLMs), where the dialogue policy planning is the key to improving the proactivity of LLMs. Most existing studies enable the dialogue policy planning of LLMs using various prompting schemes or iteratively enhance this capability in handling the given case with verbal AI feedback. However, these approaches are either bounded by the policy planning capability of the frozen LLMs or hard to be transferred to new cases. In this work, we introduce a new dialogue policy planning paradigm to strategize LLMs for proactive dialogue problems with a tunable language model plug-in as a plug-and-play dialogue policy planner, named PPDPP. Specifically, we develop a novel training framework to facilitate supervised fine-tuning over available human-annotated data as well as reinforcement learning from goal-oriented AI feedback with dynamic interaction data collected by the LLM-based self-play simulation. In this manner, the LLM-powered dialogue agent can not only be generalized to different cases after the training, but also be applicable to different applications by just substituting the learned plug-in. In addition, we propose to evaluate the policy planning capability of dialogue systems under the interactive setting. Experimental results demonstrate that PPDPP consistently and substantially outperforms existing approaches on three different proactive dialogue applications, including negotiation, emotional support, and tutoring dialogues.