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

• URL: http://arxiv.org/abs/2604.17188v2
• Date: Tue, 28 Apr 2026 11:56:12 GMT
• Title: Beyond Overlap Metrics: Rewarding Reasoning and Preferences for Faithful Multi-Role Dialogue Summarization
• Abstract summary: Multi-role dialogue summarization requires modeling complex interactions among multiple speakers.<n>Most existing methods optimize for automatic metrics such as ROUGE and BERTScore.<n>We propose a novel framework that couples explicit cognitive-style reasoning with reward-based optimization.
• Score: 5.909152930998754
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• 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 information coverage, implicit inference, factual faithfulness, and conciseness. Experiments on multilingual multi-role dialogue benchmarks show that our method matches strong baselines on ROUGE and BERTScore. Specifically, results on CSDS confirm the framework's stability in semantic consistency, while in-depth analysis on SAMSum demonstrates clear gains in factual faithfulness and model-based preference alignment. These findings underscore the value of reasoning-aware and preference-aware training for reliable dialogue summarization. Checkpoints and datasets are available at https://huggingface.co/collections/NebulaPixel/summorchestra-multirole-summary.

Related papers

• VIBE-Bench: Evaluating Personalized Large Language Models When Profiles Don't Mean Preferences [40.195544693657396]
  Existing benchmarks largely assume that preference can be retrieved from semantically related history.<n>We study an underexplored but practically important regime, profile-preference conceptual misalignment.<n>We introduce VIBE-Bench, a benchmark with two psychology-grounded tasks and 12,239 dialogues.
  arXiv  Detail & Related papers  (2026-09-01T08:45:28Z)
• Optimizing Conversational Quality in Spoken Dialogue Systems with Reinforcement Learning from AI Feedback [82.70507055599093]
  We present the first systematic study of preference learning for improving SDS quality in both multi-turn Chain-of-Thought and blockwise duplex models.<n>Experiments show that single-reward RLAIF selectively improves its targeted metric, while joint multi-reward training yields consistent gains across semantic quality and audio naturalness.
  arXiv  Detail & Related papers  (2026-01-27T00:55:14Z)
• REVEALER: Reinforcement-Guided Visual Reasoning for Element-Level Text-Image Alignment Evaluation [10.151027538362259]
  REVEALER is a unified framework for element-level alignment evaluation based on reinforcement-guided visual reasoning.<n>Our method enables Multimodal Large Language Models (MLLMs) to explicitly localize semantic elements and derive interpretable alignment judgments.
  arXiv  Detail & Related papers  (2025-12-29T03:24:09Z)
• Multi-Level Aware Preference Learning: Enhancing RLHF for Complex Multi-Instruction Tasks [81.44256822500257]
  RLHF has emerged as a predominant approach for aligning artificial intelligence systems with human preferences.<n> RLHF exhibits insufficient compliance capabilities when confronted with complex multi-instruction tasks.<n>We propose a novel Multi-level Aware Preference Learning (MAPL) framework, capable of enhancing multi-instruction capabilities.
  arXiv  Detail & Related papers  (2025-05-19T08:33:11Z)
• LLM-based Bi-level Multi-interest Learning Framework for Sequential Recommendation [54.396000434574454]
  We propose a novel multi-interest SR framework combining implicit behavioral and explicit semantic perspectives.<n>It includes two modules: the Implicit Behavioral Interest Module and the Explicit Semantic Interest Module.<n>Experiments on four real-world datasets validate the framework's effectiveness and practicality.
  arXiv  Detail & Related papers  (2024-11-14T13:00:23Z)
• Align-SLM: Textless Spoken Language Models with Reinforcement Learning from AI Feedback [50.84142264245052]
  This work introduces the Align-SLM framework to enhance the semantic understanding of textless Spoken Language Models (SLMs)<n>Our approach generates multiple speech continuations from a given prompt and uses semantic metrics to create preference data for Direct Preference Optimization (DPO)<n>We evaluate the framework using ZeroSpeech 2021 benchmarks for lexical and syntactic modeling, the spoken version of the StoryCloze dataset for semantic coherence, and other speech generation metrics, including the GPT4-o score and human evaluation.
  arXiv  Detail & Related papers  (2024-11-04T06:07:53Z)
• Rationale Behind Essay Scores: Enhancing S-LLM's Multi-Trait Essay Scoring with Rationale Generated by LLMs [2.324913904215885]
  This paper introduces Rationale-based Multiple Trait Scoring (RMTS), a novel approach for multi-trait essay scoring.<n>RMTS integrates prompt-engineering-based large language models (LLMs) with a fine-tuning-based essay scoring model using a smaller large language model (S-LLM)<n>Experiments on benchmark datasets, including ASAP, ASAP++, and Feedback Prize, show that RMTS significantly outperforms state-of-the-art models and vanilla S-LLMs in trait-specific scoring.
  arXiv  Detail & Related papers  (2024-10-18T06:35:17Z)
• CELA: Cost-Efficient Language Model Alignment for CTR Prediction [70.65910069412944]
  Click-Through Rate (CTR) prediction holds a paramount position in recommender systems.<n>Recent efforts have sought to mitigate these challenges by integrating Pre-trained Language Models (PLMs)<n>We propose textbfCost-textbfEfficient textbfLanguage Model textbfAlignment (textbfCELA) for CTR prediction.
  arXiv  Detail & Related papers  (2024-05-17T07:43:25Z)
• Hierarchical Indexing for Retrieval-Augmented Opinion Summarization [60.5923941324953]
  We propose a method for unsupervised abstractive opinion summarization that combines the attributability and scalability of extractive approaches with the coherence and fluency of Large Language Models (LLMs) Our method, HIRO, learns an index structure that maps sentences to a path through a semantically organized discrete hierarchy. At inference time, we populate the index and use it to identify and retrieve clusters of sentences containing popular opinions from input reviews.
  arXiv  Detail & Related papers  (2024-03-01T10:38:07Z)
• ED-FAITH: Evaluating Dialogue Summarization on Faithfulness [35.73012379398233]
  We first present a systematic study of faithfulness metrics for dialogue summarization. We observe that most metrics correlate poorly with human judgements despite performing well on news datasets. We propose T0-Score -- a new metric for faithfulness evaluation.
  arXiv  Detail & Related papers  (2022-11-15T19:33:50Z)
• Evaluating and Improving Factuality in Multimodal Abstractive Summarization [91.46015013816083]
  We propose CLIPBERTScore to leverage the robustness and strong factuality detection performance between image-summary and document-summary. We show that this simple combination of two metrics in the zero-shot achieves higher correlations than existing factuality metrics for document summarization. Our analysis demonstrates the robustness and high correlation of CLIPBERTScore and its components on four factuality metric-evaluation benchmarks.
  arXiv  Detail & Related papers  (2022-11-04T16:50:40Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.