Verification of Neural Network Behaviour: Formal Guarantees for Power System Applications

Abstract

This paper presents for the first time, to our knowledge, a framework for verifying neural network behavior in power system applications. Up to this moment, neural networks have been applied in power systems as a black-box; this has presented a major barrier for their adoption in practice. Developing a rigorous framework based on mixed integer linear programming, our methods can determine the range of inputs that neural networks classify as safe or unsafe, and are able to systematically identify adversarial examples. Such methods have the potential to build the missing trust of power system operators on neural networks, and unlock a series of new applications in power systems. This paper presents the framework, methods to assess and improve neural network robustness in power systems, and addresses concerns related to scalability and accuracy. We demonstrate our methods on the IEEE 9-bus, 14-bus, and 162-bus systems, treating both N-1 security and small-signal stability.

arXiv commentpublished in IEEE Transactions on Smart Grid (https://ieeexplore.ieee.org/abstract/document/9141308)

References in corpus (13)

• Explaining and Harnessing Adversarial Examples
• To prune, or not to prune: exploring the efficacy of pruning for model compression
• Efficient Neural Network Robustness Certification with General Activation Functions
• Evaluating Robustness of Neural Networks with Mixed Integer Programming
• The Power Grid Library for Benchmarking AC Optimal Power Flow Algorithms
• A Dual Approach to Scalable Verification of Deep Networks
• Data-driven Local Control Design using Optimization and Machine Learning Techniques
• A Game-Based Approximate Verification of Deep Neural Networks with Provable Guarantees
• Training for Faster Adversarial Robustness Verification via Inducing ReLU Stability
• Introducing machine learning for power system operation support
• Fast Power system security analysis with Guided Dropout
• Optimization-Based Bound Tightening using a Strengthened QC-Relaxation of the Optimal Power Flow Problem
• Provable robustness against all adversarial $l_p$-perturbations for $p\ge 1$

Cited by in corpus (5)

• Physics-Informed Neural Networks for AC Optimal Power Flow
• Neural Network Certification Informed Power System Transient Stability Preventive Control with Renewable Energy
• Neural Networks for Encoding Dynamic Security-Constrained Optimal Power Flow
• Verifying Quantized Neural Networks using SMT-Based Model Checking
• Efficient and Robust Mixed-Integer Optimization Methods for Training Binarized Deep Neural Networks