一级黄色片免费播放|中国黄色视频播放片|日本三级a|可以直接考播黄片影视免费一级毛片

高級搜索

留言板

尊敬的讀者、作者、審稿人, 關(guān)于本刊的投稿、審稿、編輯和出版的任何問題, 您可以本頁添加留言。我們將盡快給您答復(fù)。謝謝您的支持!

姓名
郵箱
手機(jī)號碼
標(biāo)題
留言內(nèi)容
驗證碼

隱私保護(hù)機(jī)器學(xué)習(xí)的密碼學(xué)方法

蔣瀚 劉怡然 宋祥福 王皓 鄭志華 徐秋亮

蔣瀚, 劉怡然, 宋祥福, 王皓, 鄭志華, 徐秋亮. 隱私保護(hù)機(jī)器學(xué)習(xí)的密碼學(xué)方法[J]. 電子與信息學(xué)報, 2020, 42(5): 1068-1078. doi: 10.11999/JEIT190887
引用本文: 蔣瀚, 劉怡然, 宋祥福, 王皓, 鄭志華, 徐秋亮. 隱私保護(hù)機(jī)器學(xué)習(xí)的密碼學(xué)方法[J]. 電子與信息學(xué)報, 2020, 42(5): 1068-1078. doi: 10.11999/JEIT190887
Han JIANG, Yiran LIU, Xiangfu SONG, Hao WANG, Zhihua ZHENG, Qiuliang XU. Cryptographic Approaches for Privacy-Preserving Machine Learning[J]. Journal of Electronics & Information Technology, 2020, 42(5): 1068-1078. doi: 10.11999/JEIT190887
Citation: Han JIANG, Yiran LIU, Xiangfu SONG, Hao WANG, Zhihua ZHENG, Qiuliang XU. Cryptographic Approaches for Privacy-Preserving Machine Learning[J]. Journal of Electronics & Information Technology, 2020, 42(5): 1068-1078. doi: 10.11999/JEIT190887

隱私保護(hù)機(jī)器學(xué)習(xí)的密碼學(xué)方法

doi: 10.11999/JEIT190887
基金項目: 國家自然科學(xué)基金(61632020, 61572294);山東省自然科學(xué)基金(ZR2017MF021);山東省科技重大創(chuàng)新工程項目(2018CXGC0702);山東半島國家自主創(chuàng)新示范區(qū)發(fā)展建設(shè)項目(S190101010001)
詳細(xì)信息
    作者簡介:

    蔣瀚:男,1974年生,講師,研究方向為密碼學(xué)與信息安全

    劉怡然:女,1996年生,博士生,研究方向為密碼學(xué)與信息安全

    宋祥福:男,1992年生,博士生,研究方向為密碼學(xué)與信息安全

    王皓:男,1984年生,副教授,研究方向為密碼學(xué)與信息安全

    鄭志華:女,1962年生,副教授,研究方向為密碼學(xué)與信息安全

    徐秋亮:男,1960年生,教授,研究方向為密碼學(xué)與信息安全

    通訊作者:

    徐秋亮 xql@sdu.edu.cn

  • 中圖分類號: TN918; TP309

Cryptographic Approaches for Privacy-Preserving Machine Learning

Funds: The National Natural Science Foundation of China (61632020, 61572294); The Natural Science Foundation of Shandong Province (ZR2017MF021); The Major Innovation Project of Science and Technology of Shandong Province (2018CXGC0702); The Funds Project of National Independent Innovation Demonstration Zone in Shandong Peninsula (S190101010001)
  • 摘要: 新一代人工智能技術(shù)的特征,表現(xiàn)為借助GPU計算、云計算等高性能分布式計算能力,使用以深度學(xué)習(xí)算法為代表的機(jī)器學(xué)習(xí)算法,在大數(shù)據(jù)上進(jìn)行學(xué)習(xí)訓(xùn)練,來模擬、延伸和擴(kuò)展人的智能。不同數(shù)據(jù)來源、不同的計算物理位置,使得目前的機(jī)器學(xué)習(xí)面臨嚴(yán)重的隱私泄露問題,因此隱私保護(hù)機(jī)器學(xué)習(xí)(PPM)成為目前廣受關(guān)注的研究領(lǐng)域。采用密碼學(xué)工具來解決機(jī)器學(xué)習(xí)中的隱私問題,是隱私保護(hù)機(jī)器學(xué)習(xí)重要的技術(shù)。該文介紹隱私保護(hù)機(jī)器學(xué)習(xí)中常用的密碼學(xué)工具,包括通用安全多方計算(SMPC)、隱私保護(hù)集合運算、同態(tài)加密(HE)等,以及應(yīng)用它們來解決機(jī)器學(xué)習(xí)中數(shù)據(jù)整理、模型訓(xùn)練、模型測試、數(shù)據(jù)預(yù)測等各個階段中存在的隱私保護(hù)問題的研究方法與研究現(xiàn)狀。
  • 圖  1  機(jī)器學(xué)習(xí)的一般過程

    圖  2  Sigmoid函數(shù)與分段激活函數(shù)圖像

    圖  3  聯(lián)邦學(xué)習(xí)

    圖  4  安全數(shù)據(jù)聚合協(xié)議

    表  1  MiniONN效率實驗結(jié)果

    數(shù)據(jù)集MNISTCIFAR10
    精確度(%)99.5291.5
    運行時間(s)32011686
    數(shù)據(jù)傳輸(MB)336.71803
    #p/h1638402524
    下載: 導(dǎo)出CSV

    表  2  文獻(xiàn)[42]效率分析

    通信計算存儲
    用戶端$O\left( {\left( {\lambda + \mu } \right)m + nr} \right)$$O\left( {m{ {\left( {\lg\left( m \right)} \right)}^2} + \left( {l + 1} \right) \cdot c\left( {r,n} \right)} \right)$$O\left( {4k\lambda + \mu \left( {m + 3\left\lceil {\dfrac{l}{2} } \right\rceil } \right) + nr} \right)$
    服務(wù)器端$O\left( { {m^2}\mu + nmr + \dfrac{ {ml} }{2} } \right)$$O\left( {{m^2} + \left( {m - \zeta } \right) \cdot l \cdot c\left( {r,n} \right)} \right)$$O\left( {mnr + {m^2}\mu + \dfrac{ {ml\mu } }{2} } \right)ght)$
    下載: 導(dǎo)出CSV
  • YAO A C. Protocols for secure computations[C]. The 23rd Annual Symposium on Foundations of Computer Science, Chicago, USA, 1982: 160–164.
    YAO A C C. How to generate and exchange secrets[C]. The 27th Annual Symposium on Foundations of Computer Science, Toronto, Canada, 1986: 162–167.
    GOLDREICH O, MICALI S, and WIGDERSON A. How to play ANY mental game[C]. The 19th Annual ACM Symposium on Theory of Computing, New York, USA, 1987: 218–229.
    BEN-OR M, GOLDWASSER S, and WIGDERSON A. Completeness theorems for non-cryptographic fault-tolerant distributed computation[C]. The 20th Annual ACM Symposium on Theory of Computing, Chicago, USA, 1988: 1–10.
    蔣瀚, 徐秋亮. 實用安全多方計算協(xié)議關(guān)鍵技術(shù)研究進(jìn)展[J]. 計算機(jī)研究與發(fā)展, 2015, 52(10): 2247–2257. doi: 10.7544/issn1000-1239.2015.20150763

    JIANG Han and XU Qiuliang. Advances in key techniques of practical secure multi-party computation[J]. Journal of Computer Research and Development, 2015, 52(10): 2247–2257. doi: 10.7544/issn1000-1239.2015.20150763
    BEAVER D. Efficient multiparty protocols using circuit randomization[C]. Annual International Cryptology Conference, Santa Barbara, USA, 1992: 420–432.
    BENDLIN R, DAMG?RD I, ORLANDI C, et al. Semi-homomorphic encryption and multiparty computation[C]. The 30th Annual International Conference on the Theory and Applications of Cryptographic Techniques, Tallinn, Estonia, 2011: 169–188.
    DAMG?RD I, PASTRO V, SMART N, et al. Multiparty computation from somewhat homomorphic encryption[C]. The 32nd Annual International Cryptology Conference, Santa Barbara, USA, 2012: 643–662.
    PAILLIER P. Public-key cryptosystems based on composite degree residuosity classes[C]. International Conference on the Theory and Applications of Cryptographic Techniques, Prague, Czech Republic, 1999: 223–238.
    GENTRY C. A fully homomorphic encryption scheme[D]. [Ph.D. dissertation], Stanford University, 2009.
    VAN DIJK M, GENTRY C, HALEVI S, et al. Fully homomorphic encryption over the integers[C]. The 29th Annual International Conference on the Theory and Applications of Cryptographic Techniques, French Riviera, France, 2010: 24–43.
    BRAKERSKI Z, GENTRY C, and VAIKUNTANATHAN V. (Leveled) fully homomorphic encryption without bootstrapping[J]. ACM Transactions on Computation Theory, 2014, 6(3): No.13.
    DUCAS L and MICCIANCIO D. FHEW: Bootstrapping homomorphic encryption in less than a second[C]. The 34th Annual International Conference on the Theory and Applications of Cryptographic Techniques, Sofia, Bulgaria, 2015: 617–640.
    孫茂華, 胡磊, 朱洪亮, 等. 布爾電路上保護(hù)隱私集合并集運算的研究與實現(xiàn)[J]. 電子與信息學(xué)報, 2016, 38(6): 1412–1418. doi: 10.11999/JEIT150911

    SUN Maohua, HU Lei, ZHU Hongliang, et al. Research and implementation of privacy preserving set union in Boolean circuits[J]. Journal of Electronics &Information Technology, 2016, 38(6): 1412–1418. doi: 10.11999/JEIT150911
    KOLESNIKOV V, KUMARESAN R, ROSULEK M, et al. Efficient batched oblivious PRF with applications to private set intersection[C]. 2016 ACM SIGSAC Conference on Computer and Communications Security, Vienna, Austria, 2016: 818–829.
    KOLESNIKOV V, ROSULEK M, TRIEU N, et al. Scalable private set union from symmetric-key techniques[EB/OL]. https://eprint.iacr.org/2019/776, 2019.
    BARNI M, ORLANDI C, and PIVA A. A privacy-preserving protocol for neural-network-based computation[C]. The 8th Workshop on Multimedia and Security, Geneva, Switzerland, 2006: 146–151.
    ORLANDI C, PIVA A, and BARNI M. Oblivious neural network computing via homomorphic encryption[J]. EURASIP Journal on Information Security, 2007, 2007(1): 037343. doi: 10.1186/1687-417X-2007-037343
    GRAEPEL T, LAUTER K, and NAEHRIG M. ML confidential: Machine learning on encrypted data[C]. The 15th International Conference on Information Security and Cryptology, Seoul, Korea, 2012: 1–21.
    ZHANG Qingchen, YANG L T, and CHEN Zhikui. Privacy preserving deep computation model on cloud for big data feature learning[J]. IEEE Transactions on Computers, 2016, 65(5): 1351–1362. doi: 10.1109/TC.2015.2470255
    HESAMIFARD E, TAKABI H, GHASEMI M, et al. Privacy-preserving machine learning in cloud[C]. 2017 ACM Cloud Computing Security Workshop, Dallas, USA, 2017: 39–43.
    ROUHANI B D, RIAZI M S, and KOUSHANFAR F. DeepSecure: Scalable provably-secure deep learning[C]. The 55th ACM/ESDA/IEEE Design Automation Conference, San Francisco, USA, 2018: 1–6.
    NIKOLAENKO V, WEINSBERG U, IOANNIDIS S, et al. Privacy-preserving ridge regression on hundreds of millions of records[C]. 2013 IEEE Symposium on Security and Privacy, Berkeley, USA, 2013: 334–348.
    MOHASSEL P and ZHANG Yupeng. SecureML: A system for scalable privacy-preserving machine learning[C]. 2017 IEEE Symposium on Security and Privacy, San Jose, USA, 2017: 19–38.
    CHANDRAN N, GUPTA D, RASTOGI A, et al. EzPC: Programmable, efficient, and scalable secure two-party computation for machine learning[EB/OL]. https://eprint.iacr.org/2017/1109, 2017.
    DEMMLER D, SCHNEIDER T, and ZOHNER M. ABY-A framework for efficient mixed-protocol secure two-party computation[C]. 2015 Network and Distributed System Security, San Diego, USA, 2015: 1–15.
    XIE Pengtao, BILENKO M, FINLEY T, et al. Crypto-nets: Neural networks over encrypted data[EB/OL]. https://arxiv.org/abs/1412.6181, 2014.
    DOWLIN N, GILAD-BACHRACH R, LAINE K, et al. CryptoNets: Applying neural networks to encrypted data with high throughput and accuracy[C]. The 33rd International Conference on Machine Learning, New York, USA, 2016: 201–210.
    BOS J W, LAUTER K, LOFTUS J, et al. Improved security for a ring-based fully homomorphic encryption scheme[C]. The 14th IMA International Conference on Cryptography and Coding, Oxford, England, 2013: 45–64.
    CHABANNE H, DE WARGNY A, MILGRAM J, et al. Privacy-preserving classification on deep neural network[EB/OL]. https://eprint.iacr.org/2017/035, 2017.
    HESAMIFARD E, TAKABI H, and GHASEMI M. CryptoDL: Deep neural networks over encrypted data[EB/OL]. https://arxiv.org/abs/1711.05189, 2017.
    LIU Jian, JUUTI M, LU Yao, et al. Oblivious neural network predictions via MiniONN transformations[C]. 2017 ACM SIGSAC Conference on Computer and Communications Security, Dallas, USA, 2017: 619–631.
    COURBARIAUX M, HUBARA I, SOUDRY D, et al. Binarized neural networks: Training deep neural networks with weights and activations constrained to +1 or –1[EB/OL]. https://arxiv.org/abs/1602.02830, 2016.
    BOURSE F, MINELLI M, MINIHOLD M, et al. Fast homomorphic evaluation of deep discretized neural networks[C]. The 38th Annual International Cryptology Conference, Santa Barbara, USA, 2018: 483–512.
    SANYAL A, KUSNER M J, GASCÓN A, et al. TAPAS: Tricks to accelerate (encrypted) prediction as a service[EB/OL]. https://arxiv.org/abs/1806.03461, 2018.
    SADEGH RIAZI M, SAMRAGH M, CHEN Hao, et al. XONN: XNOR-based oblivious deep neural network inference[C]. The 28th USENIX Conference on Security Symposium, Santa Clara, USA, 2019: 1501–1518.
    MCMAHAN H B, MOORE E, RAMAGE D, et al. Communication-efficient learning of deep networks from decentralized data[C]. The 20th International Conference on Artificial Intelligence and Statistics, Fort Lauderdale, USA, 2017: 1272–1282.
    KONEČNÝ J, MCMAHAN B, and RAMAGE D. Federated optimization: Distributed optimization beyond the datacenter[EB/OL]. https://arxiv.org/abs/1511.03575, 2015,
    KONEČNÝ J, MCMAHAN H B, YU F X, et al. Federated learning: Strategies for improving communication efficiency[EB/OL]. https://arxiv.org/abs/1610.05492, 2016.
    楊立君, 丁超, 吳蒙. 一種同時保障隱私性與完整性的無線傳感器網(wǎng)絡(luò)可恢復(fù)數(shù)據(jù)聚合方案[J]. 電子與信息學(xué)報, 2015, 37(12): 2808–2814. doi: 10.11999/JEIT150208

    YANG Lijun, DING Chao, and WU Meng. A recoverable privacy-preserving integrity-assured data aggregation scheme for wireless sensor networks[J]. Journal of Electronics &Information Technology, 2015, 37(12): 2808–2814. doi: 10.11999/JEIT150208
    BONAWITZ K, IVANOV V, KREUTER B, et al. Practical secure aggregation for privacy-preserving machine learning[C]. 2017 ACM SIGSAC Conference on Computer and Communications Security, Dallas, USA, 2017: 1175–1191.
    MANDAL K, GONG Guang, and LIU Chuyi. NIKE-based fast privacy-preserving high-dimensional data aggregation for mobile devices[R]. CACR Technical Report, CACR 2018–10, 2018.
    MANDAL K and GONG Guang. PrivFL: Practical privacy-preserving federated regressions on high-dimensional data over mobile networks[EB/OL]. https://eprint.iacr.org/2019/979, 2019.
  • 加載中
圖(4) / 表(2)
計量
  • 文章訪問數(shù):  6132
  • HTML全文瀏覽量:  3887
  • PDF下載量:  520
  • 被引次數(shù): 0
出版歷程
  • 收稿日期:  2019-11-06
  • 修回日期:  2020-03-08
  • 網(wǎng)絡(luò)出版日期:  2020-04-03
  • 刊出日期:  2020-06-04

目錄

    /

    返回文章
    返回