aiqbench analysis
This research introduces PSAP, a structured pruning technique that enhances the reliability of neural networks operating under homomorphic encryption by targeting fault-tolerant regions. By jointly optimizing weight magnitude and polynomial activation sensitivity, the method significantly reduces catastrophic accuracy drops during bit-flip injections while lowering computational overhead for encrypted inference. The work bridges critical gaps between AI model efficiency and cryptographic security constraints in privacy-preserving machine learning scenarios.
PRISM: Sensitivity-Aware PolynoMial PRuning for EffIcient Neural Network Encryption
arXiv:2607.18342v1 Announce Type: cross Abstract: Structured pruning is essential for making neural network inference feasible under homomorphic encryption (HE), yet its impact on model reliability has remained unexplored. This paper presents a systematic reliability characterization of pruned CKKS-encrypted neural networks and introduces Polynomial-Sensitivity-Aware Pruning (PSAP), a structured pruning method that is inherently reliability-aware. PSAP scores filters jointly by weight magnitude, polynomial activation sensitivity, and rotation cost, which concentrates pruning in fault-tolerant regions. Across two architectures, two datasets, two numerical representations, and five bit-error rates (40 full-model and 108 per-layer experiments), PSAP-pruned models limit catastrophic (>10 pp accuracy drop) layers to at most two versus 5--14 for magnitude-pruned baselines, reducing worst-case vulnerability by up to 29 times under int32 bit-flip injection. Direct CKKS encrypted fault injection indicates a safe operating boundary near BER~ 10^{-5}, supporting int32 injection as a conservative reliability proxy. The fault-critical structural layers account for only 1.1% of parameters, enabling selective hardening at minimal overhead. These reliability gains are obtained alongside competitive efficiency: PSAP reduces Halevi--Shoup rotations by up to 45.2\% on ResNet-32, and an adaptive mixed-degree allocation scheme lowers multiplicative depth from 66 to 56 levels, enabling leveled inference without bootstrapping.
Abstract:Structured pruning is essential for making neural network inference feasible under homomorphic encryption (HE), yet its impact on model reliability has remained unexplored. This paper presents a systematic reliability characterization of pruned CKKS-encrypted neural networks and introduces Polynomial-Sensitivity-Aware Pruning (PSAP), a structured pruning method that is inherently reliability-aware. PSAP scores filters jointly by weight magnitude, polynomial activation sensitivity, and rotation cost, which concentrates pruning in fault-tolerant regions. Across two architectures, two datasets, two numerical representations, and five bit-error rates (40 full-model and 108 per-layer experiments), PSAP-pruned models limit catastrophic (>10 pp accuracy drop) layers to at most two versus 5--14 for magnitude-pruned baselines, reducing worst-case vulnerability by up to 29 times under int32 bit-flip injection. Direct CKKS encrypted fault injection indicates a safe operating boundary near BER~ 10^{-5}, supporting int32 injection as a conservative reliability proxy. The fault-critical structural layers account for only 1.1% of parameters, enabling selective hardening at minimal overhead. These reliability gains are obtained alongside competitive efficiency: PSAP reduces Halevi--Shoup rotations by up to 45.2\% on ResNet-32, and an adaptive mixed-degree allocation scheme lowers multiplicative depth from 66 to 56 levels, enabling leveled inference without bootstrapping.
Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Hardware Architecture (cs.AR); Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG)
arXiv:2607.18342 [cs.CR]
(or arXiv:2607.18342v1 [cs.CR] for this version)
https://doi.org/10.48550/arXiv.2607.18342 arXiv-issued DOI via DataCite (pending registration)
Full text under the publisher's open licence · 223 words
Related coverage
News
Quantum Computing Report · Jul 25, 2026
DegreeSpar: Structured Degree Sparsity for Efficient Secure Transformer Inference
arXiv · Sep 30, 2026
[AINews] High Taste Testers Ep. 1
Latent Space · Oct 17, 2026
Curvature-aware dynamic precision approach for physics-informed neural networks
arXiv · Jun 5, 2026