arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.18900cs.CRcs.CV
收藏

PrivacyBench: 混合隐私保护视觉系统中的隐私并非免费

PrivacyBench: Privacy Isn't Free in Hybrid Privacy-Preserving Vision Systems

ML Collective(ML集体);Friedrich-Alexander-Universität Erlangen-Nürnberg(埃朗根-纽伦堡弗里德里希-亚历山大大学);Carnegie Mellon University(卡内基梅隆大学)

Nnaemeka Obiefuna, Samuel Oyeneye, Similoluwa Odunaiya, Iremide Oyelaja, Steven Kolawole

AI总结:

提出PrivacyBench框架,系统评估FL+DP、FL+SMPC等混合隐私技术在医学图像分类中的交互效应,发现FL+DP组合导致严重收敛失败和性能下降,而FL+SMPC保持近基线性能,为隐私-效用-成本权衡提供系统评估平台。

摘要:

在敏感深度学习应用(从医学影像到自主系统)中,隐私保护机器学习部署日益需要结合多种技术。然而,从业者缺乏系统指导来评估这些混合配置的协同和非加性交互,而是依赖孤立的技术分析,忽略了关键的系统级交互。我们引入PrivacyBench,一个基准测试框架,揭示了隐私技术组合中具有严重部署影响的显著失败。通过在医学数据集上对ResNet18和ViT模型进行系统评估,我们发现FL+DP组合表现出严重的收敛失败,准确率从98%下降到13%,同时计算成本和能耗大幅增加。相比之下,FL+SMPC以适度的开销保持了接近基线的性能。我们的框架通过自动YAML配置、资源监控和可重复的实验协议,提供了首个评估隐私-效用-成本权衡的系统平台。PrivacyBench使从业者能够在部署前识别有问题的技术交互,将隐私保护计算机视觉从临时评估推向原则性系统设计。这些发现表明隐私技术不能任意组合,并为资源受限环境中的稳健部署提供了关键指导。

英文摘要:

Privacy preserving machine learning deployments in sensitive deep learning applications; from medical imaging to autonomous systems; increasingly require combining multiple techniques. Yet, practitioners lack systematic guidance to assess the synergistic and non-additive interactions of these hybrid configurations, relying instead on isolated technique analysis that misses critical system level interactions. We introduce PrivacyBench, a benchmarking framework that reveals striking failures in privacy technique combinations with severe deployment implications. Through systematic evaluation across ResNet18 and ViT models on medical datasets, we uncover that FL + DP combinations exhibit severe convergence failure, with accuracy dropping from 98% to 13% while compute costs and energy consumption substantially increase. In contrast, FL + SMPC maintains near-baseline performance with modest overhead. Our framework provides the first systematic platform for evaluating privacy-utility-cost trade-offs through automated YAML configuration, resource monitoring, and reproducible experimental protocols. PrivacyBench enables practitioners to identify problematic technique interactions before deployment, moving privacy-preserving computer vision from ad-hoc evaluation toward principled systems design. These findings demonstrate that privacy techniques cannot be composed arbitrarily and provide critical guidance for robust deployment in resource-constrained environments.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑