发表机构
Télécom Paris, Institut Polytechnique de Paris; University of Genoa; Istituto Italiano di Tecnologia（巴黎理工学院电信学院; 热那亚大学; 意大利技术研究院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
CutClean是一种感知隐私的神经网络剪枝方法，通过辅助线性隐私头量化并消除私有属性泄露，在保持分类精度的同时实现高稀疏度，减少隐私信息流。
AI 中文摘要
神经网络越来越多地被部署在高风险应用中，同时隐私泄露问题日益受到关注。研究表明，即使不存在导致传统数据集偏差的表征不平衡，这种隐私泄露也可能发生，这在处理敏感属性的模型部署时会带来重大隐私风险。在此背景下，我们提出了CutClean，一种感知隐私的剪枝方法，该方法可减少网络中的隐私信息流，同时提高网络的稀疏性。我们的方法在每个网络块处放置辅助线性隐私头以量化信息泄露，并通过逐步提高稀疏度来消除私有属性泄露，该泄露通过附加在最后一个块上的隐私头的精度来衡量。在合成数据集和真实世界数据集上的实验表明，我们的方法可有效最小化私有信息流，同时实现高稀疏率并保持分类目标精度。
英文摘要
Neural networks are increasingly deployed in high-stakes applications with growing privacy leakage concerns. We show that this privacy leakage can occur even in the absence of representation imbalances that lead to traditional dataset biases. This poses significant privacy risks when deploying models that process sensitive attributes. In this context, we propose CutClean, a privacy-aware pruning method that allows to reduce privacy information flow through the network, while increasing its sparsity. Our approach employs auxiliary linear privacy heads placed at each network's block to quantify information leakage, and further applies increasing levels of sparsity to remove the private attribute leakage, measured in terms of the accuracy of the privacy head attached to the last block. Experiments on synthetic and real-world datasets demonstrate that our approach effectively minimizes private information flow while achieving high sparsity rates and preserving classification target accuracy.
Journal refIn: De Marsico, M., et al. Pattern Recognition. ICPR 2026. Lecture Notes in Computer Science, vol 16822. Springer, Cham
DOI：10.1007/978-3-032-31452-9_30
