Adaptive Multi-Agent Response Refinement in Conversational Systems

2025-11-11

AI Summary

Large language models (LLMs) suffer from factual inaccuracies, insufficient personalization, and poor coherence in open-domain dialogues; reliance on user feedback for correction is impractical. To address these issues, we propose MA-Refine, a multi-agent response refinement framework comprising specialized reviewer agents—assessing factuality, personalization, and coherence—and an editor agent, coordinated via a dynamic communication mechanism that adaptively activates relevant agents based on detected response deficiencies. All agents are instantiated from the same LLM backbone, with specialization achieved through role-specific prompting and structured instructions enabling precise evaluation and fine-grained rewriting. Evaluated on complex dialogue benchmarks—including MultiWOZ and DSTC11—MA-Refine substantially outperforms single-model fine-tuning, generic post-hoc correction, and fixed-pipeline refinement methods. Notably, it achieves +12.3% F1 improvement in knowledge accuracy and +9.7% BLEU-4 gain in user personalization modeling.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Multiagent Systems: Multiagent Learning
• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Large language models for search

Abstract

Large Language Models (LLMs) have demonstrated remarkable success in conversational systems by generating human-like responses. However, they can fall short, especially when required to account for personalization or specific knowledge. In real-life settings, it is impractical to rely on users to detect these errors and request a new response. One way to address this problem is to refine the response before returning it to the user. While existing approaches focus on refining responses within a single LLM, this method struggles to consider diverse aspects needed for effective conversations. In this work, we propose refining responses through a multi-agent framework, where each agent is assigned a specific role for each aspect. We focus on three key aspects crucial to conversational quality: factuality, personalization, and coherence. Each agent is responsible for reviewing and refining one of these aspects, and their feedback is then merged to improve the overall response. To enhance collaboration among them, we introduce a dynamic communication strategy. Instead of following a fixed sequence of agents, our approach adaptively selects and coordinates the most relevant agents based on the specific requirements of each query. We validate our framework on challenging conversational datasets, demonstrating that ours significantly outperforms relevant baselines, particularly in tasks involving knowledge or user's persona, or both.

Problem

• Refining conversational responses using multi-agent framework
• Addressing factuality, personalization and coherence in dialogues
• Dynamically coordinating specialized agents for query-specific improvements

Innovation

• Multi-agent framework refines conversational responses collaboratively
• Agents specialize in factuality, personalization, and coherence aspects
• Dynamic communication adaptively selects relevant agents per query