Spatio-Temporal Graph Unlearning

2025-11-12

AI Summary

To address the challenge of complying with GDPR/CCPA requirements—specifically, the need to completely erase node-level data in spatiotemporal graph models without triggering near-full retraining due to cross-spatiotemporal information propagation—this paper proposes CallosumNet. Inspired by the brain’s corpus callosum, CallosumNet integrates three key components: virtual ganglion-guided adaptive subgraph partitioning, cross-spatiotemporal dependency reconstruction, and localized forgetting updates, enabling efficient, lossless “divide-and-conquer” forgetting. Its core innovation lies in the first-ever balanced treatment of local node deletion and global structural dependency recovery in spatiotemporal graphs, avoiding disruptive global parameter perturbations. Evaluated on four real-world dynamic graph datasets from traffic and healthcare domains, CallosumNet achieves post-forgetting MAE increases of only 1–2%, closely matching full retraining performance and substantially outperforming existing state-of-the-art methods.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Spatio-temporal graphs are widely used in modeling complex dynamic processes such as traffic forecasting, molecular dynamics, and healthcare monitoring. Recently, stringent privacy regulations such as GDPR and CCPA have introduced significant new challenges for existing spatio-temporal graph models, requiring complete unlearning of unauthorized data. Since each node in a spatio-temporal graph diffuses information globally across both spatial and temporal dimensions, existing unlearning methods primarily designed for static graphs and localized data removal cannot efficiently erase a single node without incurring costs nearly equivalent to full model retraining. Therefore, an effective approach for complete spatio-temporal graph unlearning is a pressing need. To address this, we propose CallosumNet, a divide-and-conquer spatio-temporal graph unlearning framework inspired by the corpus callosum structure that facilitates communication between the brain's two hemispheres. CallosumNet incorporates two novel techniques: (1) Enhanced Subgraph Construction (ESC), which adaptively constructs multiple localized subgraphs based on several factors, including biologically-inspired virtual ganglions; and (2) Global Ganglion Bridging (GGB), which reconstructs global spatio-temporal dependencies from these localized subgraphs, effectively restoring the full graph representation. Empirical results on four diverse real-world datasets show that CallosumNet achieves complete unlearning with only 1%-2% relative MAE loss compared to the gold model, significantly outperforming state-of-the-art baselines. Ablation studies verify the effectiveness of both proposed techniques.

Problem

• Spatio-temporal graph models face GDPR unlearning challenges
• Existing methods cannot efficiently erase single node data
• Complete spatio-temporal graph unlearning requires new approaches

Innovation

• Divide-and-conquer framework for spatio-temporal graph unlearning
• Enhanced Subgraph Construction with virtual ganglions
• Global Ganglion Bridging restores full graph dependencies