Personalized Subgraph Federated Learning with Differentiable Auxiliary Projections

2025-05-29

AI Summary

To address the challenges of model alignment and aggregation arising from non-IID client subgraphs in graph federated learning, this paper proposes FedAux. The method introduces learnable Auxiliary Projection Vectors (APVs) that map local node embeddings into a one-dimensional ordered space, enabling personalized model alignment without sharing raw data or embeddings. It further employs differentiable soft sorting and 1D convolution to capture individual node characteristics, and designs a similarity-weighted parameter mixing mechanism for aggregation. Notably, FedAux is the first framework in this domain to provide a rigorous convergence analysis grounded in distributed optimization theory. Extensive experiments on multiple benchmark graph datasets demonstrate that FedAux consistently outperforms state-of-the-art methods in both global accuracy and personalized performance.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Search and Optimization: Distributed Search

Application Category

• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Federated learning (FL) on graph-structured data typically faces non-IID challenges, particularly in scenarios where each client holds a distinct subgraph sampled from a global graph. In this paper, we introduce Federated learning with Auxiliary projections (FedAux), a personalized subgraph FL framework that learns to align, compare, and aggregate heterogeneously distributed local models without sharing raw data or node embeddings. In FedAux, each client jointly trains (i) a local GNN and (ii) a learnable auxiliary projection vector (APV) that differentiably projects node embeddings onto a 1D space. A soft-sorting operation followed by a lightweight 1D convolution refines these embeddings in the ordered space, enabling the APV to effectively capture client-specific information. After local training, these APVs serve as compact signatures that the server uses to compute inter-client similarities and perform similarity-weighted parameter mixing, yielding personalized models while preserving cross-client knowledge transfer. Moreover, we provide rigorous theoretical analysis to establish the convergence and rationality of our design. Empirical evaluations across diverse graph benchmarks demonstrate that FedAux substantially outperforms existing baselines in both accuracy and personalization performance.

Problem

• Addresses non-IID challenges in federated learning on graph data
• Aligns and aggregates heterogeneous local models without sharing raw data
• Uses auxiliary projections to enable personalized cross-client knowledge transfer

Innovation

• Uses auxiliary projection vectors for alignment
• Applies soft-sorting and 1D convolution
• Computes client similarities via compact signatures