基于Rényi差分隐私的图卷积协同过滤推荐算法
作者:
中图分类号:

TP18

基金项目:

国家自然科学基金(62272077); 重庆市自然科学基金 (cstc2021jcyj-msxmX0557); 教育部人文社科规划项目(20YJAZH102)


Graph Convolutional Collaborative Filtering Recommendation Algorithm Based on Rényi Differential Privacy
Author:
  • 摘要
  • | |
  • 访问统计
  • |
  • 参考文献 [38]
  • |
  • 相似文献 [20]
  • | | |
  • 文章评论
    摘要:

    近年来, 图卷积网络作为一种强大的图嵌入技术在推荐系统领域得到广泛应用. 主要原因是推荐系统中大多数信息可以建模为图结构, 而图卷积网络是一种基于图结构的深度学习模型, 有助于挖掘图数据中用户和项目之间的潜在交互, 从而提高推荐系统的性能. 由于推荐系统的建模通常需要收集和处理大量的敏感数据, 因此可能会面临隐私泄露的风险. 差分隐私是一种具有坚实理论基础的隐私保护模型, 已被广泛应用于推荐系统中解决用户隐私泄露的问题. 目前基于差分隐私的研究主要是面向独立同分布的数据模型. 然而, 在基于图卷积网络的推荐系统中, 数据之间关联性强且不具有独立性, 这使得现有方法难以对其进行有效的隐私保护处理. 为解决该问题, 提出基于Rényi差分隐私的图卷积协同过滤推荐算法RDP-GCF, 旨在保护用户与项目交互数据安全的前提下, 实现隐私性和效用性之间的平衡. 该算法首先利用图卷积网络学习用户/项目的嵌入向量; 然后, 采用高斯机制对嵌入向量进行随机化处理, 同时基于采样的方法放大隐私预算, 减少差分噪声注入量, 以提升推荐系统的性能; 最后, 通过加权融合的方式得到用户/项目的最终嵌入向量, 并应用于推荐任务. 在3组公开数据集上进行实验验证. 结果表明, 与现有同类方法相比, 所提算法能更好地实现隐私保护与数据效用之间的平衡.

    Abstract:

    Recently, graph convolutional network (GCN), as a powerful graph embedding technology, has been widely applied in the field of recommendation. The main reason is that most of the information in recommender systems can be modeled as graph-structured data, and GCN, as a deep learning model that operates on graph structures, helps to explore the potential interactions between users and items in graph-structured data, to enhance the performance of the recommender systems. Since the modeling of recommender systems usually needs to collect and process a large amount of sensitive data, it may face the risk of privacy leakage. Differential privacy, as a privacy protection model with a solid theoretical foundation, has been widely used in recommender systems to solve the problem of personal privacy leakage. Currently, the research based on differential privacy is mainly oriented to independent and identically distributed data models. However, data within GCN-based recommender systems is highly correlated and not independent, making the existing privacy protection methods less effective. To solve the problem, this study proposes a graph convolutional collaborative filtering recommendation algorithm based on Rényi differential privacy (RDP-GCF for short), aiming to achieve a balance between privacy protection and utility while ensuring the security ofuser-item interaction data. The algorithm first utilizes GCN techniques to learn the embedding vectors for users and items. Then, the Gaussian mechanism is used to randomize the embedding vectors, and a sampling-based method is used to amplify the privacy budget and minimize the injection of differential noise, thereby improving the performance of the recommender system. Lastly, the final embedding vectors of the users and items are obtained by a weighted fusion and applied to the recommendation tasks. The proposed algorithm is validated through experiments on three publicly available datasets. The results show that compared to existing similar methods, the proposed algorithm more effectively achieves a balance between privacy protection and data utility.

    参考文献
    [1] 葛尧, 陈松灿. 面向推荐系统的图卷积网络. 软件学报, 2020, 31(4): 1101–1112. http://www.jos.org.cn/1000-9825/5928.htm
    Ge Y, Chen SC. Graph convolutional network for recommender systems. Ruan Jian Xue Bao/Journal of Software, 2020, 31(4): 1101–1112 (in Chinese with English abstract). http://www.jos.org.cn/1000-9825/5928.htm
    [2] Wang X, He XN, Wang M, Feng FL, Chua TS. Neural graph collaborative filtering. In: Proc. of the 42nd Int’l ACM SIGIR Conf. on Research and Development in Information Retrieval. Paris: ACM, 2019. 165–174.
    [3] Zhang MQ, Wu S, Yu XL, Liu Q, Wang L. Dynamic graph neural networks for sequential recommendation. IEEE Trans. on Knowledge and Data Engineering, 2023, 35(5): 4741–4753.
    [4] He XN, Deng K, Wang X, Li Y, Zhang YD, Wang M. LightGCN: Simplifying and powering graph convolution network for recommendation. In: Proc. of the 43rd Int’l ACM SIGIR Conf. on Research and Development in Information Retrieval. New York: ACM, 2020. 639–648.
    [5] Ying R, He RN, Chen KF, Eksombatchai P, Hamilton WL, Leskovec J. Graph convolutional neural networks for Web-scale recommender systems. In: Proc. of the 24th ACM SIGKDD Int’l Conf. on Knowledge Discovery & Data Mining. London: ACM, 2018. 974–983. [doi: 10.1145/3219819.3219890]
    [6] Liu Y, Yang SS, Xu YH, Miao CY, Wu M, Zhang JY. Contextualized graph attention network for recommendation with item knowledge graph. IEEE Trans. on Knowledge and Data Engineering, 2023, 35(1): 181–195.
    [7] Dai EY, Zhao TX, Zhu HS, Xu JJ, Guo ZM, Liu H, Tang JL, Wang SH. A comprehensive survey on trustworthy graph neural networks: Privacy, robustness, fairness, and explainability. arXiv:2204.08570, 2022.
    [8] Jiang HL, Pei J, Yu DX, Yu JG, Gong B, Cheng XZ. Applications of differential privacy in social network analysis: A survey. IEEE Trans. on Knowledge and Data Engineering, 2023, 35(1): 108–127.
    [9] 刘宇涵, 陈红, 刘艺璇, 赵丹, 李翠平. 图数据上的隐私攻击与防御技术. 计算机学报, 2022, 45(4): 702–734.
    Liu YH, Chen H, Liu YX, Zhao D, Li CP. State-of-the-art privacy attacks and defenses on graphs. Chinese Journal of Computers, 2022, 45(4): 702–734 (in Chinese with English abstract).
    [10] Duddu V, Boutet A, Shejwalkar V. Quantifying privacy leakage in graph embedding. In: Proc. of the 17th MobiQuitous EAI Int’l Conf. on Mobile and Ubiquitous Systems: Computing, Networking and Services. Darmstadt: ACM, 2020. 76–85.
    [11] Zhang ZK, Chen M, Backes M, Shen Y, Zhang Y. Inference attacks against graph neural networks. In: Proc. of the 31st USENIX Security Symp. Boston: USENIX Association, 2022. 4543–4560.
    [12] Wu F, Long YH, Zhang C, Li B. LINKTELLER: Recovering private edges from graph neural networks via influence analysis. In: Proc. of the 2022 IEEE Symp. on Security and Privacy (SP). San Francisco: IEEE, 2022. 2005–2024. [doi: 10.1109/SP46214.2022.9833806]
    [13] He XL, Jia JY, Backes M, Gong NZ, Zhang Y. Stealing links from graph neural networks. In: Proc. of the 30th USENIX Security Symp. Berkeley: USENIX Association, 2021. 2669–2686.
    [14] Dwork C, Roth A. The algorithmic foundations of differential privacy. Foundations and Trends® in Theoretical Computer Science, 2014, 9(3–4): 211–407.
    [15] Xie LY, Lin KX, Wang S, Wang F, Zhou JY. Differentially private generative adversarial network. arXiv:1802.06739, 2018.
    [16] 鲜征征, 李启良, 黄晓宇, 吕威, 陆寄远. 基于差分隐私和SVD++的协同过滤算法. 控制与决策, 2019, 34(1): 43–54.
    Xian ZZ, Li QL, Huang XY, Lv W, Lu JY. Collaborative filtering via SVD++ with differential privacy. Control and Decision, 2019, 34(1): 43–54 (in Chinese with English abstract).
    [17] Adesuyi TA, Kim BM. A layer-wise perturbation based privacy preserving deep neural networks. In: Proc. of the 2019 Int’l Conf. on Artificial Intelligence in Information and Communication (ICAIIC). Okinawa: IEEE, 2019. 389–394.
    [18] Hua JY, Xia C, Zhong S. Differentially private matrix factorization. In: Proc. of the 24th Int’l Conf. on Artificial Intelligence. Buenos: AAAI Press, 2015. 1763–1770.
    [19] Papernot N, Abadi M, Erlingsson Ú, Goodfellow IJ, Talwar K. Semi-supervised knowledge transfer for deep learning from private training data. arXiv:1610.05755, 2016.
    [20] 吴振强, 胡静, 田堉攀, 史武超, 颜军. 社交网络下的不确定图隐私保护算法. 软件学报, 2019, 30(4): 1106–1120. http://www.jos.org.cn/1000-9825/5368.htm
    Wu ZQ, Hu J, Tian YP, Shi WC, Yan J. Privacy preserving algorithms of uncertain graphs in social networks. Ruan Jian Xue Bao/Journal of Software, 2019, 30(4): 1106–1120 (in Chinese with English abstract). http://www.jos.org.cn/1000-9825/5368.htm
    [21] Olatunji IE, Funke T, Khosla M. Releasing graph neural networks with differential privacy guarantees. arXiv:2109.08907, 2021.
    [22] Daigavane A, Madan G, Sinha A, Thakurta AG, Aggarwal G, Jain P. Node-level differentially private graph neural networks. arXiv:2111.15521, 2021.
    [23] Abadi M, Chu A, Goodfellow I, McMahan HB, Mironov I, Talwar K, Zhang L. Deep learning with differential privacy. In: Proc. of the 2016 ACM SIGSAC Conf. on Computer and Communications Security. Vienna: ACM, 2016. 308–318. [doi: 10.1145/2976749.2978318]
    [24] Mironov I. Rényi differential privacy. In: Proc. of the 30th IEEE Computer Security Foundations Symp. (CSF). Santa Barbara: IEEE, 2017. 263–275. [doi: 10.1109/CSF.2017.11]
    [25] Liu ZQ, Wang YX, Smola A. Fast differentially private matrix factorization. In: Proc. of the 9th ACM Conf. on Recommender Systems. Vienna: ACM, 2015. 171–178. [doi: 10.1145/2792838.2800191]
    [26] Shin H, Kim S, Shin J, Xiao XK. Privacy enhanced matrix factorization for recommendation with local differential privacy. IEEE Trans. on Knowledge and Data Engineering, 2018, 30(9): 1770–1782.
    [27] Zhang SJ, Yin HZ, Chen T, Huang Z, Cui LZ, Zhang XL. Graph embedding for recommendation against attribute inference attacks. In: Proc. of the 2021 Web Conf. Ljubljana: ACM, 2021. 3002–3014. [doi: 10.1145/3442381.3449813]
    [28] Wang YX, Balle B, Kasiviswanathan SP. Subsampled Rényi differential privacy and analytical moments accountant. In: Proc. of the 22nd Int’l Conf. on Artificial Intelligence and Statistics. Naha: AISTATS, 2019. 1226–1235.
    [29] Balle B, Barthe G, Gaboardi M. Privacy amplification by subsampling: Tight analyses via couplings and divergences. In: Proc. of the 32nd Int’l Conf. on Neural Information Processing Systems. Montréal: Curran Associates Inc., 2018. 6280–6290.
    [30] Mironov I, Talwar K, Zhang L. Rényi differential privacy of the sampled Gaussian mechanism. arXiv:1908.10530, 2019.
    [31] Rendle S, Freudenthaler C, Gantner Z, Schmidt-Thieme L. BPR: Bayesian personalized ranking from implicit feedback. In: Proc. of the 25th Conf. on Uncertainty in Artificial Intelligence. Montreal: AUAI Press, 2009. 452-461.
    [32] Kingma DP, Ba JL. Adam: A method for stochastic optimization. In: Proc. of the 3rd Int’l Conf. on Learning Representations. San Diego: ICLR, 2015.
    [33] He XN, Liao LZ, Zhang HW, Nie LQ, Hu X, Chua TS. Neural collaborative filtering. In: Proc. of the 26th Int’l Conf. on World Wide Web. Perth: Int’l World Wide Web Conf. Steering Committee, 2017. 173–182. [doi: 10.1145/3038912.3052569]
    [34] Glorot X, Bengio Y. Understanding the difficulty of training deep feedforward neural networks. In: Proc. of the 13th Int’l Conf. on Artificial Intelligence and Statistics. Sardinia: PMLR, 2010. 249–256.
    引证文献
    网友评论
    网友评论
    分享到微博
    发 布
引用本文

王锟,王永,刘金源,邓江洲.基于Rényi差分隐私的图卷积协同过滤推荐算法.软件学报,2025,36(3):1202-1217

复制
分享
文章指标
  • 点击次数:365
  • 下载次数: 794
  • HTML阅读次数: 200
  • 引用次数: 0
历史
  • 收稿日期:2023-07-19
  • 最后修改日期:2023-10-02
  • 在线发布日期: 2024-11-01
文章二维码
您是第20059224位访问者
版权所有:中国科学院软件研究所 京ICP备05046678号-3
地址:北京市海淀区中关村南四街4号,邮政编码:100190
电话:010-62562563 传真:010-62562533 Email:jos@iscas.ac.cn
技术支持:北京勤云科技发展有限公司

京公网安备 11040202500063号