MediHive: A Decentralized Agent Collective for Medical Reasoning

2026-03-28

AI Summary

This work addresses the limitations of single-agent systems in handling uncertainty and conflicting evidence in complex medical reasoning, as well as the poor scalability, single-point failure risks, and role ambiguity inherent in centralized multi-agent architectures. To overcome these challenges, the authors propose a decentralized multi-agent framework grounded in large language models, featuring a shared memory pool, a conditional evidence-based debate mechanism, and a multi-round local information fusion algorithm. This design enables agents to autonomously assign roles, collaboratively analyze evidence, and iteratively reach consensus without centralized coordination. The approach marks the first effective application of a decentralized multi-agent system to high-stakes medical reasoning, achieving state-of-the-art accuracy of 84.3% on MedQA and 78.4% on PubMedQA—significantly outperforming both single-agent and centralized baselines—while enhancing robustness, scalability, and fault tolerance.

Abstract

Large language models (LLMs) have revolutionized medical reasoning tasks, yet single-agent systems often falter on complex, interdisciplinary problems requiring robust handling of uncertainty and conflicting evidence. Multi-agent systems (MAS) leveraging LLMs enable collaborative intelligence, but prevailing centralized architectures suffer from scalability bottlenecks, single points of failure, and role confusion in resource-constrained environments. Decentralized MAS (D-MAS) promise enhanced autonomy and resilience via peer-to-peer interactions, but their application to high-stakes healthcare domains remains underexplored. We introduce MediHive, a novel decentralized multi-agent framework for medical question answering that integrates a shared memory pool with iterative fusion mechanisms. MediHive deploys LLM-based agents that autonomously self-assign specialized roles, conduct initial analyses, detect divergences through conditional evidence-based debates, and locally fuse peer insights over multiple rounds to achieve consensus. Empirically, MediHive outperforms single-LLM and centralized baselines on MedQA and PubMedQA datasets, attaining accuracies of 84.3% and 78.4%, respectively. Our work advances scalable, fault-tolerant D-MAS for medical AI, addressing key limitations of centralized designs while demonstrating superior performance in reasoning-intensive tasks.

Problem

• medical reasoning
• multi-agent systems
• decentralized architecture
• large language models
• healthcare AI

Innovation

• Decentralized Multi-Agent System
• Medical Reasoning
• Shared Memory Pool
• Iterative Fusion
• LLM-based Agents