CM2: Multimodal Cultural Reasoning via an Integrated Multi-Agent Framework
2026-08-31 • Artificial Intelligence
Artificial Intelligence
AI summaryⓘ
The authors present CM2, a system designed to help large language models better understand and reason about information that comes from different types of sources, like text and images, in a way similar to how humans interpret culture. CM2 uses multiple connected agents and combines perception, retrieval, reasoning, and feedback to improve the model's ability to make decisions across different modalities. Their tests show that CM2 performs better than existing step-by-step reasoning methods, and each part of the system contributes to this improvement. They also demonstrate that CM2 can effectively resolve conflicts between information from different sources.
Multimodal Large Language ModelsCultural ReasoningRetrieval-Augmented GenerationNetworked ReasoningGated FusionReward-Driven FeedbackChain-of-Thought ReasoningMultimodal PerceptionCross-Modal ArbitrationAblation Study
Authors
Qi Li, Zhaojie Kang, Yingjie He, Zheng Lin, Hao Zhang, Guangxin Wu, Yan Gong, Rong Fu, Jianyuan Ni
Abstract
Multimodal Large Language Models (MLLMs) have shown remarkable success in STEM domains, where progress is often driven by vertical, step-by-step deduction under relatively stable symbol systems. Their horizontal, interdisciplinary cultural reasoning, however, remains underexplored.We propose CM2, a multi-agent framework grounded in the cognitive pathway of human cultural interpretation. CM2 integrates multimodal perception, retrieval-augmented generation, networked reasoning, gated fusion, and reward-driven feedback.Experiments on CM2D across multiple MLLM backbones show consistent gains over CoT and typical reasoning paradigms; ablations validate each module's contribution, and conflict analyses confirm genuine cross-modal arbitration.