Collaborative training of neural networks leverages distributed data by exchanging gradient information between different clients. Although training data entirely resides with the clients, recent work shows that training data can be reconstructed from such exchanged gradient information. To enhance privacy, gradient perturbation techniques have been proposed. However, they come at the cost of reduced model performance, increased convergence time, or increased data demand. In this paper, we introduce PRECODE, a PRivacy EnhanCing mODulE that can be used as generic extension for arbitrary model architectures. We propose a simple yet effective realization of PRECODE using variational modeling. The stochastic sampling induced by variational modeling effectively prevents privacy leakage from gradients and in turn preserves privacy of data owners. We evaluate PRECODE using state of the art gradient inversion attacks on two different model architectures trained on three datasets. In contrast to commonly used defense mechanisms, we find that our proposed modification consistently reduces the attack success rate to 0% while having almost no negative impact on model training and final performance. As a result, PRECODE reveals a promising path towards privacy enhancing model extensions.
ThinkEval: Practical Evaluation of Knowledge Leakage in LLM Editing using Thought-based Knowledge Graphs
Electrical and Computer Engineering National University of Singapore（新加坡国立大学电子与计算机工程系）；A*STAR Institute for Infocomm Research (A*STAR I 2 R), Singapore（新加坡A*STAR信息与通信研究机构）
随着如今数据的规模和大小显著增加，用于高性能计算和机器学习的联邦学习（Bonawitz et al.[2019]）比以往任何时候都更加重要（Abadi et al.[2016]）。人们过去认为，共享梯度似乎是安全的，以隐藏在训练阶段的本地训练数据。然而，Zhu等人[2019]证明了可以通过检测梯度从模型训练数据中恢复原始数据。它们使用生成的随机虚拟数据，并最小化它们与真实数据之间的距离。Zhao et al.[2020]进一步推动了收敛算法。通过用交叉熵损失代替原有的损失函数，获得了更好的保真度阈值。在本文中，我们建议在将数据馈送到分类器之前使用额外的dropout（Srivastava et al.[2014]）层。它在防止原始数据泄漏方面非常有效，因为即使在5，800个epoch之后，训练数据也不能收敛到小的RMSE，其中丢弃率设置为0.5。
Real-time Fuel Leakage Detection via Online Change Point Detection
School of computing technologies（计算技术学院）；RMIT university（皇家墨尔本理工大学）；Titan Cloud Software（泰坦云软件公司）；School of Computer and Mathematical Sciences（计算机与数学科学学院）；University of Adelaide（阿德莱德大学）
ISACL: Internal State Analyzer for Copyrighted Training Data Leakage
City University of Hong Kong（香港城市大学）；Microsoft（微软）；University of Illinois Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）；Google LLC（谷歌公司）；Arizona State University（亚利桑那州立大学）；Stevens Institute of Technology（史蒂文斯理工学院）
BridgePure: Limited Protection Leakage Can Break Black-Box Data Protection
Academy of Mathematics and Systems Science, Chinese Academy of Sciences（中国科学院数学与系统科学研究院）；University of Chinese Academy of Sciences（中国科学院大学）；University of Waterloo（滑铁卢大学）；Vector Institute（向量研究所）
University of Kalyani（卡利亚尼大学）；National Institute of Science, Education and Research Bhubaneswar（布巴内斯瓦尔国家科学教育与研究学院）；Homi Bhaba National Institute（霍米·巴巴国家学院）
Encoded Gradients Aggregation against Gradient Leakage in Federated Learning
University of Electronic Science and Technology of China（电子科技大学）；Peng Cheng Lab（鹏城实验室）；Shenzhen Research Institute, The Chinese University of Hong Kong（香港中文大学深圳研究院）；The Chinese University of Hong Kong（香港中文大学）；Harbin Institute of Technology, Shenzhen（哈尔滨工业大学（深圳））
