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

2026-03-16

AI Summary

This work addresses the challenge of deploying deep neural networks on resource-constrained embedded systems, where compression techniques such as quantization and pruning are essential but may compromise behavioral consistency with the original model—particularly in safety-critical applications. To tackle this issue, the paper proposes a probabilistic certification framework that supports heterogeneous compressed architectures through a dual-network symbolic propagation mechanism. By integrating variance-aware bound analysis based on Bernstein inequalities, the method enables tight verification of behavioral similarity under adjustable confidence levels. The approach significantly improves both the tightness and scalability of certification, outperforming existing techniques on benchmarks including ACAS Xu and computer vision tasks, thereby offering efficient and reliable quantitative safety guarantees for compressed models.

Technology Category

• Machine Learning: Learning on the Edge & Model Compression
• Constraint Satisfaction and Optimization: Satisfiability Modulo Theories
• Natural Language Processing: Safety and Robustness

Application Category

• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Security and Privacy: Large-scale security measurements
• Systems and Infrastructure for Web, Mobile and WoT: Web performance, measurement, and characterization

Abstract

Deploying Deep Neural Networks (DNNs) on resource-constrained embedded systems requires aggressive model compression techniques like quantization and pruning. However, ensuring that the compressed model preserves the behavioral fidelity of the original design is a critical challenge in the safety-critical system design flow. Existing verification methods often lack scalability or fail to handle the architectural heterogeneity introduced by pruning. In this work, we propose SimCert, a probabilistic certification framework for verifying the behavioral similarity of compressed neural networks. Unlike worst-case analysis, SimCert provides quantitative safety guarantees with adjustable confidence levels. Our framework features: (1) A dual-network symbolic propagation method supporting both quantization and pruning; (2) A variance-aware bounding technique using Bernstein's inequality to tighten safety certificates; and (3) An automated verification toolchain. Experimental results on ACAS Xu and computer vision benchmarks demonstrate that SimCert outperforms state-of-the-art baselines.

Problem

• behavioral similarity
• neural network compression
• safety-critical systems
• model verification
• quantization and pruning

Innovation

• probabilistic certification
• behavioral similarity
• neural network compression
• symbolic propagation
• variance-aware bounding