graph-unlearning

IsleNet introduces a spatial-entropy-based partitioning method for spatiotemporal graph unlearning, enabling exact data removal with low computational cost while maintaining high accuracy for privacy regulations.

The paper proposes MMFGU, a multimodal federated graph unlearning framework that decouples target-specific representations to handle entity, modality, and pairing removal requests while preserving retained utility, achieving a 41.5x speedup over full retraining.