Physics-Aware Machine Unlearning for Cyber-Physical Systems

2026-09-28

AI Summary

This study addresses the challenge that machine unlearning in cyber-physical systems (CPS) often violates physical constraints and compromises safe deployment. To this end, we propose a physics-aware machine unlearning method that couples physical residuals as safety guardrails into the gradient ascent process, guiding weight updates toward the feasible region and preventing convergence to arbitrary surrogate solutions. By integrating a Physics-Informed Neural Network (PINN) controller with high-fidelity OpenDSS co-simulation, the proposed approach achieves simultaneous poisoned data removal and restoration of physical compliance. Experimental evaluations on the IEEE 34-node test feeder demonstrate that the method effectively eliminates poisoned data while reconstructing physical feasibility, significantly outperforming existing baselines.

Abstract

This paper proposes a physics-guided gradient-ascent-based machine unlearning method that couples the forgetting signal with the physical residual of the target cyber-physical systems, ensuring that weight updates during unlearning are steered toward physically feasible regions of the weight space. The physics residual acts as a safety fence during gradient ascent: the model is steered away from the poisoned behavioral basin and simultaneously toward physics-compliant territory, rather than toward an arbitrary alternative that may still violate domain constraints. We evaluate the proposed method against four baselines: naive gradient ascent, exact unlearning, SISA, and full retraining on an IEEE 34-bus distribution system, driven by two physics-informed neural network-based distribution energy resource controllers and validated through high-fidelity OpenDSS power-flow co-simulation. From the evaluation, we found that our proposed physics-guided model simultaneously removes poison and restores physical compliance, which are essential for the safe deployment of safety-critical cyber-physical systems

Problem

• Machine Unlearning
• Cyber-Physical Systems
• Physics Compliance
• Safety-Critical Systems
• Data Poisoning

Innovation

• Machine Unlearning
• Physics-Informed Neural Networks
• Cyber-Physical Systems
• Gradient Ascent
• Physical Residual