Spectral Embeddings Leak Graph Topology: Theory, Benchmark, and Adaptive Reconstruction

TL;DR

This paper introduces a comprehensive framework for understanding and reconstructing fragmented, noisy, and privacy-leaking graph data using spectral methods, with theoretical guarantees and practical benchmarks.

Contribution

It proposes LoGraB for benchmarking fragmented graphs and AFR for adaptive spectral reconstruction, providing new theoretical insights and state-of-the-art results.

Findings

AFR achieves the best F1 on 7 out of 9 datasets.

LoGraB reveals model strengths and weaknesses under fragmentation.

AFR retains 75% of its undefended F1 under differential privacy at epsilon=2.

Abstract

Graph Neural Networks (GNNs) excel on relational data, but standard benchmarks unrealistically assume the graph is centrally available. In practice, settings such as Federated Graph Learning, distributed systems, and privacy-sensitive applications involve graph data that are localized, fragmented, noisy, and privacy-leaking. We present a unified framework for this setting. We introduce LoGraB (Local Graph Benchmark), which decomposes standard datasets into fragmented benchmarks using three strategies and four controls: neighborhood radius $d$, spectral quality $k$, noise level $\sigma$, and coverage ratio $p$. LoGraB supports graph reconstruction, localized node classification, and inter-fragment link prediction, with Island Cohesion. We propose AFR (Adaptive Fidelity-driven Reconstruction), a method for noisy spectral fragments. AFR scores patch quality via a fidelity measure combining a gap-to-truncation stability ratio and structural entropy, then assembles fragments using RANSAC-Procrustes alignment, adaptive stitching, and Bundle Adjustment. Rather than forcing a single global graph, AFR recovers large faithful islands. We prove heat-kernel edge recovery under a separation condition, Davis--Kahan perturbation stability, and bounded alignment error. We establish a Spectral Leakage Proposition: under a spectral-gap assumption, polynomial-time Bayesian recovery is feasible once enough eigenvectors are shared, complementing AFR's deterministic guarantees. Experiments on nine benchmarks show that LoGraB reveals model strengths and weaknesses under fragmentation, AFR achieves the best F1 on 7/9 datasets, and under per-embedding $(\epsilon,\delta)$-Gaussian differential privacy, AFR retains 75% of its undefended F1 at $\epsilon=2$. Our anonymous code is available at https://anonymous.4open.science/r/JMLR_submission