Efficient Reachability Analysis for Convolutional Neural Networks Using Hybrid Zonotopes
This work addresses the need for scalable and accurate verification of neural networks in autonomous systems, representing an incremental improvement over existing methods.
The paper tackled the problem of formal verification for convolutional neural networks in safety-critical applications by developing a novel set-based reachability analysis method, achieving a flexible trade-off between computational complexity and approximation accuracy.
Feedforward neural networks are widely used in autonomous systems, particularly for control and perception tasks within the system loop. However, their vulnerability to adversarial attacks necessitates formal verification before deployment in safety-critical applications. Existing set propagation-based reachability analysis methods for feedforward neural networks often struggle to achieve both scalability and accuracy. This work presents a novel set-based approach for computing the reachable sets of convolutional neural networks. The proposed method leverages a hybrid zonotope representation and an efficient neural network reduction technique, providing a flexible trade-off between computational complexity and approximation accuracy. Numerical examples are presented to demonstrate the effectiveness of the proposed approach.