Unified-MAS: Universally Generating Domain-Specific Nodes for Empowering Automatic Multi-Agent Systems

TL;DR

Unified-MAS introduces a two-stage framework for generating domain-specific nodes in multi-agent systems, improving performance and reducing costs in knowledge-intensive tasks by decoupling node creation from system orchestration.

Contribution

It proposes a novel offline node synthesis approach that enhances MAS adaptability across domains, overcoming internal knowledge limits of large language models.

Findings

Achieves up to 14.2% performance improvement

Reduces system costs significantly

Demonstrates robustness across multiple domains

Abstract

Automatic Multi-Agent Systems (MAS) generation has emerged as a promising paradigm for solving complex reasoning tasks. However, existing frameworks are fundamentally bottlenecked when applied to knowledge-intensive domains (e.g., healthcare and law). They either rely on a static library of general nodes like Chain-of-Thought, which lack specialized expertise, or attempt to generate nodes on the fly. In the latter case, the orchestrator is not only bound by its internal knowledge limits but must also simultaneously generate domain-specific logic and optimize high-level topology, leading to a severe architectural coupling that degrades overall system efficacy. To bridge this gap, we propose Unified-MAS that decouples granular node implementation from topological orchestration via offline node synthesis. Unified-MAS operates in two stages: (1) Search-Based Node Generation retrieves external open-world knowledge to synthesize specialized node blueprints, overcoming the internal knowledge limits of LLMs; and (2) Reward-Based Node Optimization utilizes a perplexity-guided reward to iteratively enhance the internal logic of bottleneck nodes. Extensive experiments across four specialized domains demonstrate that integrating Unified-MAS into four Automatic-MAS baselines yields a better performance-cost trade-off, achieving up to a 14.2% gain while significantly reducing costs. Further analysis reveals its robustness across different designer LLMs and its effectiveness on conventional tasks such as mathematical reasoning.