arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2404.09265cs.CRcs.AI
收藏

拆分而非劫持:防御拆分学习中的特征空间劫持攻击

Make Split, not Hijack: Preventing Feature-Space Hijacking Attacks in Split Learning

Tampere University(坦佩雷大学);RISE Research Institutes of Sweden(瑞典RISE研究院)

Tanveer Khan, Mindaugas Budzys, Antonis Michalas

AI总结:

本研究针对拆分学习易受特征空间劫持等攻击、隐私保护不足的问题,提出结合拆分学习与函数秘密共享的混合方法,通过随机掩码与秘密份额处理实现数据隐私保护,兼具高准确率、低通信开销与训练效率优势。

摘要:

机器学习(ML)的普及使得敏感数据的隐私保护变得前所未有的重要。拆分学习(SL)这类协作学习技术旨在在优化ML流程的同时保护客户端数据。尽管前景可观,但SL已被证明易受多种攻击,因此引发了人们对其数据隐私保护有效性的担忧。在本研究中,我们提出了一种结合SL与函数秘密共享(Function Secret Sharing, FSS)的混合方法,以保障客户端数据隐私。客户端在将激活图发送给服务器之前,会向其添加一个随机掩码。服务器无法获取原始函数,只能处理通过FSS生成的秘密份额。因此,在前向传播和反向传播过程中,服务器都无法从激活图中重构出客户端的原始数据。此外,通过视觉可逆性实验我们证明,使用FSS时服务器无法从激活图中重构原始图像数据。与服务器可获取客户端输入信息的其他基于SL的方法相比,该方法减少了隐私泄露,提升了隐私保护能力。我们的方法还能抵御特征空间劫持攻击,防止敏感信息被潜在篡改。我们的协议取得了理想效果:与仅使用FSS、不采用SL的相同模型相比,通信开销降低了2倍以上,训练时间缩短了7倍以上。同时我们还证明,该方法的准确率超过96%,与明文模型的性能相当。

英文摘要:

The popularity of Machine Learning (ML) makes the privacy of sensitive data more imperative than ever. Collaborative learning techniques like Split Learning (SL) aim to protect client data while enhancing ML processes. Though promising, SL has been proved to be vulnerable to a plethora of attacks, thus raising concerns about its effectiveness on data privacy. In this work, we introduce a hybrid approach combining SL and Function Secret Sharing (FSS) to ensure client data privacy. The client adds a random mask to the activation map before sending it to the servers. The servers cannot access the original function but instead work with shares generated using FSS. Consequently, during both forward and backward propagation, the servers cannot reconstruct the client's raw data from the activation map. Furthermore, through visual invertibility, we demonstrate that the server is incapable of reconstructing the raw image data from the activation map when using FSS. It enhances privacy by reducing privacy leakage compared to other SL-based approaches where the server can access client input information. Our approach also ensures security against feature space hijacking attack, protecting sensitive information from potential manipulation. Our protocols yield promising results, reducing communication overhead by over 2x and training time by over 7x compared to the same model with FSS, without any SL. Also, we show that our approach achieves >96% accuracy and remains equivalent to the plaintext models.

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。

↑