[2604.17188] Beyond Overlap Metrics: Rewarding Reasoning and Preferences for Faithful Multi-Role Dialogue Summarization

Computer Science > Computation and Language arXiv:2604.17188 (cs) [Submitted on 19 Apr 2026 (v1), last revised 28 Apr 2026 (this version, v2)] Title:Beyond Overlap Metrics: Rewarding Reasoning and Preferences for Faithful Multi-Role Dialogue Summarization Authors:Xiaoyong Mei, Tingting Zuo, Da Chen, Guangyu Hu, Xiangyu Wen, Chao Duan, Mingyan Zhang, Fudan Zheng View a PDF of the paper titled Beyond Overlap Metrics: Rewarding Reasoning and Preferences for Faithful Multi-Role Dialogue Summarization, by Xiaoyong Mei and 6 other authors View PDF HTML (experimental) Abstract:Multi-role dialogue summarization requires modeling complex interactions among multiple speakers while preserving role-specific information and factual consistency. However, most existing methods optimize for automatic metrics such as ROUGE and BERTScore, which favor surface-level imitation of references rather than genuine gains in faithfulness or alignment with human preferences. We propose a novel framework that couples explicit cognitive-style reasoning with reward-based optimization for multi-role dialogue summarization. Our method first distills structured reasoning traces (e.g., step-by-step inferences and intermediate reflections) from a large teacher model and uses them as auxiliary supervision to initialize a reasoning-aware summarizer via staged supervised fine-tuning. It then applies GRPO with a dual-principle reward that blends metric-based signals with human-aligned criteria targeting key info...