.png){kind=icon}
.png){kind=icon}
Return
SheafAlign: A Sheaf-Theoretic Framework for Decentralized Multimodal Alignment
.png){kind=icon}
OA
AI
DOI：10.1109/LCOMM.2026.3663044.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Conventional multimodal alignment methods assume mutual redundancy across all modalities, an assumption that fails in real-world distributed scenarios. We propose Sheaf-Align, a sheaf-theoretic framework for decentralized multimodal alignment that replaces single-space alignment with multiple comparison spaces. This approach models pairwise modality relations through sheaf structures and leverages decentralized contrastive learning-based objectives for training. SheafAlign overcomes the limitations of prior methods by not requiring mutual redundancy among all modalities, preserving both shared and unique information. Experiments on multimodal sensing datasets show superior zero-shot generalization, cross-modal alignment, and robustness to missing modalities, with 34% lower communication cost than state-of-the-art baselines.
Keywords:
Distributed optimization
distributed sensing
sheaf theory
multimodal learning
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.