全同态加密CNN推理的内存与噪声协同优化方法OA

To address the challenges of high memory consumption,low computational efficiency,and homomorphic noise accumulation in fully homomorphic encryption(FHE)for privacy-preserving inference in convolutional neural network(CNN),this paper proposes a collaborative optimization framework.The framework introduces a hierarchical memory scheduling strategy,which employs a dynamic key loading mechanism and an adaptive compression technique for polynomial ring slot numbers(reducing available slots exponentially based on network depth),thereby significantly decreasing memory usage.Additionally,a noise suppression residual module is developed,incorporating a noise propagation dynamics model to design a real-time noise monitoring-based on-demand bootstrapping trigger mechanism,which reduces bootstrapping frequency and enhances inference efficiency.Experimental results on the CIFAR-10 dataset demonstrate that this framework enables homomorphic encrypted inference of ResNet-20 in approximately 500 s with only 20 GB of memory,achieving a 3.5×improvement in inference efficiency and a 94％reduction in memory consumption compared to existing CKKS-based solutions(2 271 s/384 GB).This framework provides a novel technical paradigm for privacy-preserving machine learning in resource-constrained scenarios.