arXiv 2604.071002026-04-20cs.CLcs.AI

STRIDE-ED：一种基于策略的分步推理框架用于共情对话系统

STRIDE-ED: A Strategy-Grounded Stepwise Reasoning Framework for Empathetic Dialogue Systems

School of Artificial Intelligence, OPtics and ElectroNics (iOPEN), Northwestern Polytechnical University, China（人工智能学院、光学与电子学（iOPEN）、西北工业大学，中国）；School of Artificial Intelligence and Big Data, Henan University of Technology, China（人工智能与大数据学院、河南理工大学，中国）；School of Computer Science, Northwestern Polytechnical University, China（计算机学院、西北工业大学，中国）

Hongru Ji, Yuyin Fan, Meng Zhao, Xianghua Li, Lianwei Wu, Chao Gao

备注：: Accepted by ACL 2026

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

本文提出STRIDE-ED框架，通过结构化策略引导推理，解决共情对话中策略意识和上下文敏感决策的难题，实验表明其在多种LLM上表现优异。

摘要：

共情对话不仅需要识别用户情绪状态，还需在生成响应时做出策略意识强、上下文敏感的决策。然而，缺乏全面的共情策略框架、显式的任务对齐多阶段推理以及高质量的策略意识数据限制了现有方法，无法有效建模共情对话作为复杂多阶段认知和决策过程。为解决这些挑战，我们提出了STRIDE-ED，一种基于策略、可解释且深度推理的框架，通过结构化、策略条件的推理建模共情对话。为支持有效学习，我们开发了一种策略意识数据精炼管道，整合基于LLM的标注、多模型一致性加权评估和动态采样，构建与共情策略一致的高质量训练数据。此外，我们采用两阶段训练范式，结合监督微调与多目标强化学习，以更好地对齐模型行为与目标情绪、共情策略和响应格式。广泛实验表明，STRIDE-ED在多种开源LLM上泛化能力强，并在自动指标和人工评估中均优于现有方法。我们的数据和代码已公开在https://github.com/jicoder-nwpu/STRIDE-ED。

英文摘要：

Empathetic dialogue requires not only recognizing a user's emotional state but also making strategy-aware, context-sensitive decisions throughout response generation. However, the lack of a comprehensive empathy strategy framework, explicit task-aligned multi-stage reasoning, and high-quality strategy-aware data fundamentally limits existing approaches, preventing them from effectively modeling empathetic dialogue as a complex, multi-stage cognitive and decision-making process. To address these challenges, we propose STRIDE-ED, a STRategy-grounded, Interpretable, and DEep reasoning framework that models Empathetic Dialogue through structured, strategy-conditioned reasoning. To support effective learning, we develop a strategy-aware data refinement pipeline integrating LLM-based annotation, multi-model consistency-weighted evaluation, and dynamic sampling to construct high-quality training data aligned with empathetic strategies. Furthermore, we adopt a two-stage training paradigm that combines supervised fine-tuning with multi-objective reinforcement learning to better align model behaviors with target emotions, empathetic strategies, and response formats. Extensive experiments demonstrate that STRIDE-ED generalizes across diverse open-source LLMs and consistently outperforms existing methods on both automatic metrics and human evaluations. Our data and code are publicly available at https://github.com/jicoder-nwpu/STRIDE-ED.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。