Convert Language Model into a Value-based Strategic Planner

TL;DR

This paper introduces straQ*, a framework that transforms language models into strategic planners using Q-learning, improving long-term emotional support in conversations by optimizing responses based on future rewards.

Contribution

The paper presents a novel plug-and-play framework that enables LLMs to plan strategically for emotional support conversations using reinforcement learning.

Findings

straQ* outperforms baseline methods in ESC tasks

Q-learning enhances LLMs' ability to plan long-term strategies

Framework demonstrates significant improvements on ESC datasets

Abstract

Emotional support conversation (ESC) aims to alleviate the emotional distress of individuals through effective conversations. Although large language models (LLMs) have obtained remarkable progress on ESC, most of these studies might not define the diagram from the state model perspective, therefore providing a suboptimal solution for long-term satisfaction. To address such an issue, we leverage the Q-learning on LLMs, and propose a framework called straQ*. Our framework allows a plug-and-play LLM to bootstrap the planning during ESC, determine the optimal strategy based on long-term returns, and finally guide the LLM to response. Substantial experiments on ESC datasets suggest that straQ* outperforms many baselines, including direct inference, self-refine, chain of thought, finetuning, and finite state machines.