AgentCDM: Enhancing Multi-Agent Collaborative Decision-Making via ACH-Inspired Structured Reasoning

2025-08-16

AI Summary

Current large language model (LLM)-based multi-agent systems (MAS) for collaborative decision-making (CDM) face two key bottlenecks: authoritarian strategies are vulnerable to individual cognitive biases, while voting-based mechanisms fail to fully harness collective intelligence. To address these limitations, this paper proposes AgentCDM—a novel framework that, for the first time, integrates the cognitive science principle of Analysis of Competing Hypotheses (ACH) into MAS-based CDM. AgentCDM establishes a structured two-stage reasoning paradigm: hypothesis generation followed by critical evaluation. Furthermore, it introduces a two-stage de-scaffolding training methodology to internalize and generalize reasoning capabilities. Extensive experiments across multiple benchmark tasks demonstrate that AgentCDM significantly outperforms existing baselines in robustness, generalization, and decision quality. This work establishes a new paradigm for trustworthy, cognitively grounded collaborative decision-making in LLM-MAS.

Technology Category

• Multiagent Systems: Mechanism Design
• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Multi-agent systems (MAS) powered by large language models (LLMs) hold significant promise for solving complex decision-making tasks. However, the core process of collaborative decision-making (CDM) within these systems remains underexplored. Existing approaches often rely on either ``dictatorial" strategies that are vulnerable to the cognitive biases of a single agent, or ``voting-based" methods that fail to fully harness collective intelligence. To address these limitations, we propose extbf{AgentCDM}, a structured framework for enhancing collaborative decision-making in LLM-based multi-agent systems. Drawing inspiration from the Analysis of Competing Hypotheses (ACH) in cognitive science, AgentCDM introduces a structured reasoning paradigm that systematically mitigates cognitive biases and shifts decision-making from passive answer selection to active hypothesis evaluation and construction. To internalize this reasoning process, we develop a two-stage training paradigm: the first stage uses explicit ACH-inspired scaffolding to guide the model through structured reasoning, while the second stage progressively removes this scaffolding to encourage autonomous generalization. Experiments on multiple benchmark datasets demonstrate that AgentCDM achieves state-of-the-art performance and exhibits strong generalization, validating its effectiveness in improving the quality and robustness of collaborative decisions in MAS.

Problem

• Addresses cognitive biases in multi-agent decision-making systems
• Shifts from passive answer selection to active hypothesis evaluation
• Enhances collaborative reasoning quality and robustness in MAS

Innovation

• Structured reasoning paradigm inspired by ACH
• Two-stage training with scaffolding removal
• Active hypothesis evaluation instead of passive selection