Chain of Modality: From Static Fusion to Dynamic Orchestration in Omni-MLLMs

arXiv:2604.14520v1 Announce Type: new Abstract: Omni-modal Large Language Models (Omni-MLLMs) promise a unified integration of diverse sensory streams. However, recent evaluations reveal a critical performance paradox: unimodal baselines frequently outperform joint multimodal inference. We trace this perceptual fragility to the static fusion topologies universally employed by current models, identifying two structural pathologies: positional bias in sequential inputs and alignment traps in interleaved formats, which systematically distort attention regardless of task semantics. To resolve this functional rigidity, we propose Chain of Modality (CoM), an agentic framework that transitions multimodal fusion from passive concatenation to dynamic orchestration. CoM adaptively orchestrates input topologies, switching among parallel, sequential, and interleaved pathways to neutralize structural biases. Furthermore, CoM bifurcates cognitive execution into two task-aligned pathways: a streamlined Direct-Decide'' path for direct perception and a structured Reason-Decide'' path for analytical auditing. Operating in either a training-free or a data-efficient SFT setting, CoM achieves robust and consistent generalization across diverse benchmarks.

Related

• Reasoning Within the Mind: Dynamic Multimodal Interleaving in Latent Space
• See Fair, Speak Truth: Equitable Attention Improves Grounding and Reduces Hallucination in Vision-Language Alignment
• Why Multimodal In-Context Learning Lags Behind? Unveiling the Inner Mechanisms and Bottlenecks
• DecAlign: Hierarchical Cross-Modal Alignment for Decoupled Multimodal Representation Learning

Source: arXiv cs.CV | 2026-04-17