Graph Learning Across Data Silos

TL;DR

This paper introduces a novel federated graph learning framework that enables multiple clients to collaboratively learn personalized and consensus graphs from distributed data without sharing raw data, addressing heterogeneity and privacy concerns.

Contribution

It proposes an auto-weighted multi-graph learning model with a tailored algorithm that preserves data privacy and handles data heterogeneity, supported by theoretical guarantees.

Findings

Effective graph learning in distributed, privacy-sensitive scenarios

Theoretical bounds on estimation error and convergence

Superior performance on synthetic and real datasets

Abstract

We consider the problem of inferring graph topology from smooth graph signals in a novel but practical scenario where data are located in distributed clients and prohibited from leaving local clients due to factors such as privacy concerns. The main difficulty in this task is how to exploit the potentially heterogeneous data of all clients under data silos. To this end, we first propose an auto-weighted multiple graph learning model to jointly learn a personalized graph for each local client and a single consensus graph for all clients. The personalized graphs match local data distributions, thereby mitigating data heterogeneity, while the consensus graph captures the global information. Moreover, the model can automatically assign appropriate contribution weights to local graphs based on their similarity to the consensus graph. We next devise a tailored algorithm to solve the induced problem, where all raw data are processed locally without leaving clients. Theoretically, we establish a provable estimation error bound and convergence analysis for the proposed model and algorithm. Finally, extensive experiments on synthetic and real data are carried out, and the results illustrate that our approach can learn graphs effectively in the target scenario.