---
title: "Hybrid Federated Learning for Multimodal Fusion under Missing Modalities and Site Heterogeneity"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2611.03032v1"
permalink: "https://feed.astralor.com/items/itm_f8d2bd6dee42fb2307a2118f"
published: "2026-11-02T12:29:30.000Z"
score: 56
channel: "ai"
tags: ["多模态","Hybrid Federated Learning","Multimodal Fusion","Missing Modalities","Site Heterogeneity Multimodal","FedSMKL","Multimodal"]
---

# Hybrid Federated Learning for Multimodal Fusion under Missing Modalities and Site Heterogeneity

## At a glance

Multimodal fusion is studied under hybrid federation, where data are partitioned across samples and feature blocks. The proposed two-dimensional federated smooth multiple-kernel large-margin learning method (2D-FedSMKL)…

## Key points

- Multimodal fusion is studied under hybrid federation, where data are partitioned across samples and feature blocks.
- The proposed two-dimensional federated smooth multiple-kernel large-margin learning method (2D-FedSMKL) integrates availability-corrected modality scores, group-kernel regularization, and smooth worst-site risk aggregat…
- For fixed maps and propensities, the objective is convex and smooth.

## Full text

Multimodal fusion is studied under hybrid federation, where data are partitioned across samples and feature blocks. The proposed two-dimensional federated smooth multiple-kernel large-margin learning method (2D-FedSMKL) integrates availability-corrected modality scores, group-kernel regularization, and smooth worst-site risk aggregation. For fixed maps and propensities, the objective is convex and smooth. Curvature bounds and exact synchronous block-gradient equivalence are established for binary and multiclass scores, giving a conditional accelerated rate. An oracle-propensity binary risk bound and a propensity-perturbation result are also derived. Simulations and two public activity-recognition benchmarks assess site heterogeneity and modality absence. Worst-site aggregation improves worst-site performance under heterogeneity, including reversed and size-balanced designs. On the benchmarks, validation selects the correction; the selected model is not significantly worse than the strongest reference at the worst site, and group-regularization benefits vary by dataset.

---

Verify against the original before citing; curated by Astral Feed: https://feed.astralor.com/items/itm_f8d2bd6dee42fb2307a2118f