论文标题
放松：捍卫会员推理攻击而不会丢失实用程序
RelaxLoss: Defending Membership Inference Attacks without Losing Utility
论文作者
论文摘要
作为对培训数据隐私的长期威胁，会员推理攻击（MIA）在机器学习模型中无处不在。现有作品证明了训练的区分性与测试损失分布与模型对MIA的脆弱性之间的密切联系。在现有结果的推动下，我们提出了一个基于轻松损失的新型培训框架，其学习目标更为可实现，从而导致概括差距狭窄并减少了隐私泄漏。 RelaseLoss适用于任何分类模型，具有易于实施和可忽略不计的开销的额外好处。通过对五个具有不同方式（图像，医疗数据，交易记录）的数据集进行广泛的评估，我们的方法在针对MIAS和模型效用方面始终优于最先进的防御机制。我们的防守是第一个可以承受广泛攻击的同时，同时保存（甚至改善）目标模型的效用。源代码可从https://github.com/dingfanchen/relaxloss获得
As a long-term threat to the privacy of training data, membership inference attacks (MIAs) emerge ubiquitously in machine learning models. Existing works evidence strong connection between the distinguishability of the training and testing loss distributions and the model's vulnerability to MIAs. Motivated by existing results, we propose a novel training framework based on a relaxed loss with a more achievable learning target, which leads to narrowed generalization gap and reduced privacy leakage. RelaxLoss is applicable to any classification model with added benefits of easy implementation and negligible overhead. Through extensive evaluations on five datasets with diverse modalities (images, medical data, transaction records), our approach consistently outperforms state-of-the-art defense mechanisms in terms of resilience against MIAs as well as model utility. Our defense is the first that can withstand a wide range of attacks while preserving (or even improving) the target model's utility. Source code is available at https://github.com/DingfanChen/RelaxLoss
