Litcius/Paper detail

Certifying Geometric Robustness of Neural Networks

Mislav Balunović, Maximilian Baader, Gagandeep Singh, Timon Gehr, Martin Vechev

2020Repository for Publications and Research Data (ETH Zurich)61 citationsDOIOpen Access PDF

Abstract

The use of neural networks in safety-critical computer vision systems calls for their robustness certification against natural geometric transformations (e.g., rotation, scaling).However, current certification methods target mostly norm-based pixel perturbations and cannot certify robustness against geometric transformations.In this work, we propose a new method to compute sound and asymptotically optimal linear relaxations for any composition of transformations.Our method is based on a novel combination of sampling and optimization.We implemented the method in a system called DEEPG and demonstrated that it certifies significantly more complex geometric transformations than existing methods on both defended and undefended networks while scaling to large architectures.

Topics & Concepts

Robustness (evolution)Computer scienceArtificial neural networkTransformation geometryCertificationScalingPixelArtificial intelligenceAlgorithmTheoretical computer scienceMathematical optimizationMathematicsGeometryPolitical scienceBiochemistryLawGeneChemistryAdversarial Robustness in Machine LearningAdvanced Neural Network ApplicationsAnomaly Detection Techniques and Applications