SimCert: Probabilistic Certification for Behavioral Similarity in Deep Neural Network Compression

Researchers developed SimCert, a probabilistic certification framework that verifies behavioral similarity between compressed neural networks and their original versions. The framework addresses critical safety challenges in deploying compressed DNNs on resource-constrained systems by providing quantitative safety guarantees with adjustable confidence levels.

• →SimCert introduces a dual-network symbolic propagation method supporting both quantization and pruning for neural network compression verification.
• →The framework uses variance-aware bounding with Bernstein's inequality to provide tighter safety certificates than existing methods.
• →Unlike worst-case analysis approaches, SimCert offers probabilistic guarantees with adjustable confidence levels for safety-critical systems.
• →Experimental results on ACAS Xu and computer vision benchmarks show SimCert outperforms current state-of-the-art baselines.
• →The framework includes an automated verification toolchain to handle architectural heterogeneity in compressed neural networks.