Safety Verification and Robustness Analysis of Neural Networks via Quadratic Constraints and Semidefinite Programming

• Published in: IEEE transactions on automatic control Vol. 67; no. 1; pp. 1 - 15
• Main Authors: Fazlyab, Mahyar, Morari, Manfred, Pappas, George J.
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.01.2022; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Biological neural networks; Convex optimization; deep neural networks; Machine learning; Neural networks; Perturbation methods; Programming; Robustness; Robustness (mathematics); robustness analysis; Safety; safety verification; Semidefinite programming; semidefinite programming (SDP); Uncertainty; Verification

Certifying the safety or robustness of neural networks against input uncertainties and adversarial attacks is an emerging challenge in the area of safe machine learning and control. To provide such a guarantee, one must be able to bound the output of neural networks when their input changes within a...