MMAC-Copilot: Multi-modal Agent Collaboration Operating Copilot

TL;DR

MMAC-Copilot enhances multi-modal agent collaboration by leveraging diverse agent expertise, significantly improving interaction capabilities and reducing hallucinations across various application domains, as demonstrated on benchmark tests.

Contribution

This work introduces the Multi-Modal Agent Collaboration framework (MMAC-Copilot), enabling multi-agent teamwork to improve application interaction and reduce hallucinations in large language model agents.

Findings

Achieved 6.8% performance improvement on GAIA benchmark.

Demonstrated strong capabilities on VIBench for non-API applications.

Showcased effective multi-agent collaboration across diverse domains.

Abstract

Large language model agents that interact with PC applications often face limitations due to their singular mode of interaction with real-world environments, leading to restricted versatility and frequent hallucinations. To address this, we propose the Multi-Modal Agent Collaboration framework (MMAC-Copilot), a framework utilizes the collective expertise of diverse agents to enhance interaction ability with application. The framework introduces a team collaboration chain, enabling each participating agent to contribute insights based on their specific domain knowledge, effectively reducing the hallucination associated with knowledge domain gaps. We evaluate MMAC-Copilot using the GAIA benchmark and our newly introduced Visual Interaction Benchmark (VIBench). MMAC-Copilot achieved exceptional performance on GAIA, with an average improvement of 6.8\% over existing leading systems. VIBench focuses on non-API-interactable applications across various domains, including 3D gaming, recreation, and office scenarios. It also demonstrated remarkable capability on VIBench. We hope this work can inspire in this field and provide a more comprehensive assessment of Autonomous agents. The anonymous Github is available at \href{https://anonymous.4open.science/r/ComputerAgentWithVision-3C12}{Anonymous Github}