Lune

ACM MM2025Top-tier venue

A Theoretical Proof of Dynamic Multimodal Fusion Exacerbates Modality Greedy

Xiaorui Ding, Huan Ma, Changqing Zhang

Abstract

In recent years, numerous studies have proposed uncertainty-guided multimodal learning to adapt to dynamic relationships between different modalities. Dynamic multimodal fusion enables more dominant modalities to receive greater weight during the fusion process, thereby preventing the influence of spurious features from less reliable modalities on decision-making. However, there is No free lunch. We observe that the introduction of dynamic fusion during training exacerbates the model's tendency toward Greedy (a phenomenon known to induce decision shortcuts in multimodal learning). This results in a model that does not fully take advantage of the lower quality modalities. In this paper, we provide a theoretical analysis showing that dynamic fusion intensifies Greedy, and we present experimental results that support this observation. In summary, this paper explains the Greedy risk in dynamic multimodal learning from both theoretical and experimental perspectives, serving as a cautionary reminder for researchers when employing dynamic multimodal learning. Our code is available at https://github.com/d-xr/GreedyDynFusion.

Ask Lune

Know where this paper stands in minutes, not an afternoon.

Lune can read the top-tier papers that cite it and the ones closest to it, and tells you what it claims, what came after and where it falls short, each point cited to the paper it comes from. No skimming, no AI slop.

Questions to start from

Lune can draw on

its abstract · related work

Get the answer

Free to start. No credit card required.

Related papers

  1. Provable Dynamic Fusion for Low-Quality Multimodal DataQingyang Zhang, Haitao Wu, Changqing Zhang, Qinghua Hu et al.ICML 2023 · 143 citations
  2. Unbiased Dynamic Multimodal FusionShicai Wei, Kaijie Zhang, Luyi Chen, Tao He, Guiduo DuanCVPR Findings 2026
  3. CL-DMDF: Dynamic Multimodal Data Fusion Model Based on Contrastive LearningDong Li, Lingling Zhang, Binghao Han, Linlin Ding et al.AAAI 2026
  4. G2D: Boosting Multimodal Learning with Gradient-Guided DistillationMohammed Rakib, Arunkumar BagavathiICCV 2025 · 1 citation
  5. Predictive Dynamic FusionBing Cao, Yinan Xia, Yi Ding, Changqing Zhang et al.ICML 2024 · 31 citations
Dusk over the sea between two cliffs drawn in fine vertical lines