Toward Federated Multimodal Graph Foundation Models: A Topology-Aware Multimodal Alignment Framework
Machine Learning
Summary
The authors study graphs where nodes have multiple types of information like images and text, which are common in areas like social media and healthcare. These graphs are often spread across different places that can't share raw data, so the authors propose a method called FedGAMMA to collaboratively learn from them without data exposure. FedGAMMA works in two steps: first, it aligns the different types of data and graph structure during pre-training, and then it fine-tunes the model using special prompts for each client's data. Their experiments show FedGAMMA performs better than other methods, especially when limited labeled data is available.
Authors
Xunkai Li, Guohao Fu, Yuming Ai, Zhengyu Wu, Hongchao Qin, Rong-Hua Li, Guoren Wang
Abstract
Multimodal-attributed graphs (MAGs), whose nodes carry modalities such as images and text alongside topological structure, now pervade applications including social platforms, e-commerce, and biomedical networks, offering richer semantic signals than single-modality graphs. In practice, such graphs are fragmented across privacy-restricted silos owned by different platforms and institutions, so learning a broadly transferable model over them demands collaborative training that never exposes raw data. This places the task at the intersection of multimodal graph learning and federated learning, yet existing methods cover only one side of it. To address the challenges from these two perspectives, we propose FedGAMMA, casting federated multimodal graph foundation learning as a two-stage semantic-structural alignment problem of federated pre-training and prompt-based fine-tuning. During pre-training, a shared-private semantic enhancer disentangles cross-modal commonality from modality-specific information, aligning it through optimal transport, a topology-aware graph fusion module decouples semantic and structural views via semantic residual graphs and dual positional encodings, and a dual-channel affinity-aware aggregation mechanism estimates client similarity from feature and graph centroids without exposing raw data. During fine-tuning, FedGAMMA adapts the pretrained encoder through lightweight graph-aware prompts, a shared prompt pool with controlled exploration, and channel-wise prompt synchronization. Experiments on twelve multimodal graph datasets show FedGAMMA consistently surpassing a broad range of baselines across downstream tasks, with gains of up to 12.96%. FedGAMMA further outperforms competitive baselines accross multi-domain datasets on multiple tasks with up to 5.71% under few-shot learning scenario.