InEx: Hallucination Mitigation via Introspection and Cross-Modal Multi-Agent Collaboration

TL;DR

InEx is a training-free multi-agent framework that reduces hallucinations in multimodal large language models by combining introspective reasoning with cross-modal collaboration, leading to more reliable AI responses.

Contribution

It introduces a novel, autonomous, multi-agent approach inspired by human decision-making to effectively mitigate hallucinations without additional training.

Findings

Achieves 4%-27% improvements on hallucination benchmarks.

Demonstrates robustness across various tasks.

Utilizes entropy-based uncertainty for introspection.

Abstract

Hallucination remains a critical challenge in large language models (LLMs), hindering the development of reliable multimodal LLMs (MLLMs). Existing solutions often rely on human intervention or underutilize the agent's ability to autonomously mitigate hallucination. To address these limitations, we draw inspiration from how humans make reliable decisions in the real world. They begin with introspective reasoning to reduce uncertainty and form an initial judgment, then rely on external verification from diverse perspectives to reach a final decision. Motivated by this cognitive paradigm, we propose InEx, a training-free, multi-agent framework designed to autonomously mitigate hallucination. InEx introduces internal introspective reasoning, guided by entropy-based uncertainty estimation, to improve the reliability of the decision agent's reasoning process. The agent first generates a response, which is then iteratively verified and refined through external cross-modal multi-agent collaboration with the editing agent and self-reflection agents, further enhancing reliability and mitigating hallucination. Extensive experiments show that InEx consistently outperforms existing methods, achieving 4%-27% gains on general and hallucination benchmarks, and demonstrating strong robustness.