基于注意力机制的联邦无线流量预测模型
作者:
中图分类号:

TP393

基金项目:

国家自然科学基金(62272256, 62202250, 61832012); 山东省自然科学基金基础研究重大基础研究项目(ZR2022ZD03); 山东省自然科学基金(ZR2021QF079, ZR2023MF040); 齐鲁工业大学(山东省科学院)科教融合创新试点项目(2022XD001); 齐鲁工业大学(山东省科学院)培优基金(2023PY059); 济南“新高校20条”资助项目(20228093)


Federated Wireless Traffic Prediction Model Based on Attention Mechanism
Author:
  • 摘要
  • | |
  • 访问统计
  • |
  • 参考文献 [58]
  • |
  • 相似文献 [20]
  • | | |
  • 文章评论
    摘要:

    移动数据每天都在不断增长, 如何精准预测无线流量对高效、合理的配置通信和网络资源至关重要. 现有的流量预测方法多采用集中式训练架构, 涉及大规模的流量数据传输, 会导致用户隐私泄露等安全问题. 联邦学习可以在数据本地存储的前提下训练一个全局模型, 保护用户隐私, 有效减轻数据频繁传输负担. 但是在无线流量预测中, 单个基站数据量有限, 且不同基站流量数据模式异构, 流量模式难以捕捉, 导致训练得到的全局模型泛化能力较差. 此外, 传统联邦学习方法在进行模型聚合时采用简单平均, 忽略了客体贡献差异, 进一步导致全局模型性能下降. 针对上述问题, 提出一种基于注意力的“类内平均, 类间注意力”联邦无线流量预测模型, 该模型根据基站的流量数据进行聚类, 更好地捕捉具有相似流量模式基站的流量变化特性; 同时, 设计一个预热模型, 利用少量基站数据缓解数据异构, 提高全局模型的泛化能力; 在模型聚合阶段引入注意力机制, 量化不同客体对全局模型的贡献, 并在模型迭代过程中融入预热模型, 大幅提升模型的预测精度. 在两个真实数据集(Milano和Trento)上进行大量实验, 结果表明该方法优于所有基线方法. 并且与目前最先进的方法相比, 在两个数据集上的平均绝对误差性能增益最高分别达到10.1%和9.6%.

    Abstract:

    As mobile data is growing everyday, how to predicate the wireless traffic accurately is crucial for the efficient and sensible allocation of communication and network resources. However, most existing prediction methods use a centralized training architecture, which involves large-scale traffic data transmission, leading to security issues such as user privacy leakage. Federated learning can train a global model with local data storage, which protects users’ privacy and effectively reduces the burden of frequent data transmission. However, in wireless traffic prediction, the amount of data from the single base station is limited, and the traffic patterns vary among different base stations, making it difficult to capture the traffic patterns and resulting in poor generalization of the global model. In addition, traditional federated learning methods employ averaging in model aggregation, ignoring the differences in guest contributions, which further leads to the degradation of the global model performance. To address the above issues, this study proposes an attention-based “intra-cluster average, inter-cluster attention” federated wireless traffic prediction model. The model first clusters base stations based on their traffic data to better capture the traffic variation characteristics of base stations with similar traffic patterns. At the same time, a warm-up model is designed to alleviate data heterogeneity by a small amount of base station data to improve the generalization ability of the global model. The study introduces the attention mechanism in the aggregation stage to quantify the contributions of different objects to the global model and incorporates the warm-up model in the model iteration process to improve the prediction accuracy of the model. Extensive experiments are conducted on two real-world datasets (Milano and Trento), and the results show that the DualICA outperforms all baseline methods. The mean absolute error performance gain over the state-of-the-art method is up to 10.1% and 9.6% on the two datasets, respectively.

    参考文献
    [1] David K, Berndt H. 6G vision and requirements: Is there any need for beyond 5G? IEEE Vehicular Technology Magazine, 2018, 13(3): 72–80. [doi: 10.1109/MVT.2018.2848498]
    [2] Zong BQ, Fan C, Wang XY, Duan XY, Wang BJ, Wang JW. 6G technologies: Key drivers, core requirements, system architectures, and enabling technologies. IEEE Vehicular Technology Magazine, 2019, 14(3): 18–27.
    [3] Wang Z, Wong VWS. Bayesian meta-learning for adaptive traffic prediction in wireless networks. IEEE Trans. on Mobile Computing, 2023, 23(6): 6620–6633.
    [4] Jiang WW. Cellular traffic prediction with machine learning: A survey. Expert Systems with Applications, 2022, 201: 117163.
    [5] Niu ZS, Wu YQ, Gong J, Yang ZX. Cell zooming for cost-efficient green cellular networks. IEEE Communications Magazine, 2010, 48(11): 74–79.
    [6] Niu ZS. TANGO: Traffic-aware network planning and green operation. IEEE Wireless Communications, 2011, 18(5): 25–29.
    [7] Kato N, Mao BM, Tang FX, Kawamoto Y, Liu JJ. Ten challenges in advancing machine learning technologies toward 6G. IEEE Wireless Communications, 2020, 27(3): 96–103.
    [8] Yang Q, Liu Y, Chen TJ, Tong YX. Federated machine learning: Concept and applications. ACM Trans. on Intelligent Systems and Technology, 2019, 10(2): 12.
    [9] Li T, Sahu AK, Talwalkar A, Smith V. Federated learning: Challenges, methods, and future directions. IEEE Signal Processing Magazine, 2020, 37(3): 50–60.
    [10] Kairouz P, McMahan HB, Avent B, et al. Advances and open problems in federated learning. Foundations and Trends® in Machine Learning, 2021, 14(1–2): 1–210.
    [11] 杨强. 联邦学习: 人工智能的最后一公里. 智能系统学报, 2020, 15(1): 183–186.
    Yang Q. Federated learning: The last on kilometer of artificial intelligence. CAAI Trans. on Intelligent Systems, 2020, 15(1): 183–186 (in Chinese with English abstract).
    [12] 杨强. AI与数据隐私保护: 联邦学习的破解之道. 信息安全研究, 2019, 5(11): 961–965.
    Yang Q. AI and data privacy protection: The way to federated learning. Journal of Information Security Research, 2019, 5(11): 961–965 (in Chinese with English abstract).
    [13] Nasr M, Shokri R, Houmansadr A. Comprehensive privacy analysis of deep learning: Passive and active white-box inference attacks against centralized and federated learning. In: Proc. of the 2019 IEEE Symp. on Security and Privacy (SP). San Francisco: IEEE, 2019. 739-753. [doi: 10.1109/SP.2019.00065]
    [14] Chen H, Zhu T, Zhang T, Zhou WL, Yu PS. Privacy and fairness in federated learning: On the perspective of tradeoff. ACM Computing Surveys, 2023, 56(2): 1–37.
    [15] Shu YT, Yu MF, Yang O, Liu JK, Feng HF. Wireless traffic modeling and prediction using seasonal ARIMA models. IEICE Trans. on Communications, 2005, 88(10): 3992–3999.
    [16] Zhou B, He D, Sun Z. Traffic predictability based on ARIMA/GARCH model. In: Proc. of the 2nd Conf. on Next Generation Internet Design and Engineering. Valencia: IEEE, 2006. 207.
    [17] Cappé O, Moulines E, Pesquet JC, Petropulu AP, Yang XS. Long-range dependence and heavy-tail modeling for teletraffic data. IEEE Signal Processing Magazine, 2002, 19(3): 14–27.
    [18] Ashtiani F, Salehi JA, Aref MR. Mobility modeling and analytical solution for spatial traffic distribution in wireless multimedia networks. IEEE Journal on Selected Areas in Communications, 2003, 21(10): 1699–1709.
    [19] Tutschku K, Tran-Gia P. Spatial traffic estimation and characterization for mobile communication network design. IEEE Journal on Selected Areas in Communications, 1998, 16(5): 804–811.
    [20] Li RP, Zhao ZF, Zheng JC, Mei CL, Cai YM, Zhang HG. The learning and prediction of application-level traffic data in cellular networks. IEEE Trans. on Wireless Communications, 2017, 16(6): 3899–3912.
    [21] Zhang Y, Roughan M, Willinger W, Qiu LL. Spatio-temporal compressive sensing and Internet traffic matrices. In: Proc. of the 2009 ACM SIGCOMM Conf. on Data Communication. Barcelona: ACM, 2009. 267–278. [doi: 10.1145/1592568.1592600]
    [22] Soule A, Lakhina A, Taft N, Papagiannaki K, Salamatian K, Nucci A, Crovella M, Diot C. Traffic matrices: Balancing measurements, inference and modeling. In: Proc. of the 2005 ACM SIGMETRICS Int’l Conf. on Measurement and Modeling of Computer Systems. Banff: ACM, 2005. 362–373. [doi: 10.1145/1064212.1064259]
    [23] Sharma S, Majumdar A, Elvira V, Chouzenoux E. Blind Kalman filtering for short-term load forecasting. IEEE Trans. on Power Systems, 2020, 35(6): 4916–4919.
    [24] Li RP, Zhao ZF, Wei Y, Zhou X, Zhang HG. GM-PAB: A grid-based energy saving scheme with predicted traffic load guidance for cellular networks. In: Proc. of the 2012 IEEE Int’l Conf. on Communications (ICC). Ottawa: IEEE, 2012. 1160–1164.
    [25] Li RP, Zhao ZF, Zhou X, Palicot J, Zhang HG. The prediction analysis of cellular radio access network traffic: From entropy theory to networking practice. IEEE Communications Magazine, 2014, 52(6): 234–240.
    [26] Chen XM, Jin YH, Qiang SW, Hu WS, Jiang KD. Analyzing and modeling spatio-temporal dependence of cellular traffic at city scale. In: Proc. of the 2015 IEEE Int’l Conf. on Communications (ICC). London: IEEE, 2015. 3585–3591. [doi: 10.1109/ICC.2015.7248881]
    [27] Liu H, Xu B, Lu DJ, Zhang GJ. A path planning approach for crowd evacuation in buildings based on improved artificial bee colony algorithm. Applied Soft Computing, 2018, 68: 360–376.
    [28] Ghahramani Z. Probabilistic machine learning and artificial intelligence. Nature, 2015, 521(7553): 452–459.
    [29] Su XG, Yan X, Tsai CL. Linear regression. WIREs Computational Statistics, 2012, 4(3): 275–294.
    [30] Sapankevych NI, Sankar R. Time series prediction using support vector machines: A survey. IEEE Computational Intelligence Magazine, 2009, 4(2): 24–38.
    [31] Feng HF, Shu YT, Wang SY, Ma MD. SVM-based models for predicting WLAN traffic. In: Proc. of the 2006 IEEE Int’l Conf. on Communications. Istanbul: IEEE, 2006. 597–602. [doi: 10.1109/ICC.2006.254860]
    [32] Zuo C, Qian JM, Feng SJ, Yin W, Li YX, Fan PF, Han J, Qian KM, Chen Q. Deep learning in optical metrology: A review. Light: Science & Applications, 2022, 11(1): 39.
    [33] Kato N, Fadlullah ZM, Mao BM, Tang FX, Akashi O, Inoue T, Mizutani K. The deep learning vision for heterogeneous network traffic control: Proposal, challenges, and future perspective. IEEE Wireless Communications, 2017, 24(3): 146–153.
    [34] Mao Q, Hu F, Hao Q. Deep learning for intelligent wireless networks: A comprehensive survey. IEEE Communications Surveys & Tutorials, 2018, 20(4): 2595–2621.
    [35] Wang J, Tang J, Xu ZY, Wang YZ, Xue GL, Zhang X, Yang DJ. Spatiotemporal modeling and prediction in cellular networks: A big data enabled deep learning approach. In: Proc. of the 2017 IEEE Conf. on Computer Communications (INFOCOM 2017). Atlanta: IEEE, 2017. 1–9.
    [36] Qiu C, Zhang YY, Feng ZY, Zhang P, Cui SG. Spatio-temporal wireless traffic prediction with recurrent neural network. IEEE Wireless Communications Letters, 2018, 7(4): 554–557.
    [37] Zhang CT, Zhang HX, Qiao JP, Yuan DF, Zhang MG. Deep transfer learning for intelligent cellular traffic prediction based on cross-domain big data. IEEE Journal on Selected Areas in Communications, 2019, 37(6): 1389–1401.
    [38] Wu Q, He KW, Chen X, Yu S, Zhang JS. Deep transfer learning across cities for mobile traffic prediction. IEEE/ACM Trans. on Networking, 2022, 30(3): 1255–1267.
    [39] Lin CY, Su HT, Tung SL, Hsu WH. Multivariate and propagation graph attention network for spatial-temporal prediction with outdoor cellular traffic. In: Proc. of the 30th ACM Int’l Conf. on Information & Knowledge Management. ACM, 2021. 3248–3252.
    [40] Yao Y, Gu B, Su Z, Guizani M. MVSTGN: A multi-view spatial-temporal graph network for cellular traffic prediction. IEEE Trans. on Mobile Computing, 2023, 22(5): 2837–2849
    [41] Bonawitz K, Ivanov V, Kreuter B, Marcedone A, McMahan HB, Patel S, Ramage D, Segal A, Seth K. Practical secure aggregation for privacy-preserving machine learning. In: Proc. of the 2017 ACM SIGSAC Conf. on Computer and Communications Security. Dallas: ACM, 2017. 1175-1191. [doi: 10.1145/3133956.3133982]
    [42] McMahan B, Moore E, Ramage D, Hampson S, Arcas BAY. Communication-efficient learning of deep networks from decentralized data. In: Proc. of the 20th Int’l Conf. on Artificial Intelligence and Statistics. Fort Lauderdale: PMLR, 2017. 1273–1282.
    [43] Zhao Y, Li M, Lai LZ, Suda N, Civin D, Chandra V. Federated learning with Non-IID data. arXiv:1806.00582, 2018.
    [44] Li T, Sahu K, Zaheer M, Sanjabi M, Talwalkar A, Smith V. Federated optimization in heterogeneous networks. In: Proc. of the 2020 Machine Learning and Systems. Austin: MLSys, 2020. 429–450.
    [45] Ji SX, Pan SR, Long GD, Li X, Jiang J, Huang Z. Learning private neural language modeling with attentive aggregation. In: Proc. of the 2019 Int’l Joint Conf. on Neural Networks (IJCNN). Budapest: IEEE, 2019. 1–8. [doi: 10.1109/IJCNN.2019.8852464]
    [46] Shu JG, Yang TT, Liao YX, Chen FR, Xiao Y, Yang K, Jia XH. Clustered federated multitask learning on Non-IID data with enhanced privacy. IEEE Internet of Things Journal, 2023, 10(4): 3453–3467.
    [47] Yang L, Huang Jm, Lin Wy, Cao JN. Personalized federated learning on Non-IID data via group-based meta-learning. ACM Trans. on Knowledge Discovery from Data, 2023, 17(4): 49.
    [48] Arisdakessian S, Wahab OA, Mourad A, Otrok H. Coalitional federated learning: Improving communication and training on Non-IID data with selfish clients. IEEE Trans. on Services Computing, 2023, 16(4): 2462–2476.
    [49] Sun QH, Li X, Zhang JY, Xiong L, Liu WR, Liu JF, Qin Z, Ren K. ShapleyFL: Robust federated learning based on shapley value. In: Proc. of the 29th ACM SIGKDD Conf. on Knowledge Discovery and Data Mining. Long Beach: ACM, 2023. 2096–2108.
    [50] Mai WM, Yao JC, Chen G, Zhang Y, Cheung YM, Han B. Server-client collaborative distillation for federated reinforcement learning. ACM Trans. on Knowledge Discovery from Data, 2023, 18(1): 9.
    [51] Zhang CT, Dang SP, Shihada B, Alouini MS. Dual attention-based federated learning for wireless traffic prediction. In: Proc. of the 2021 IEEE Conf. on Computer Communications (INFOCOM 2021). Vancouver: IEEE, 2021. 1–10. [doi: 10.1109/INFOCOM42981.2021.9488883]
    [52] Zheng TH, Li BC. Poisoning attacks on deep learning based wireless traffic prediction. In: Proc. of the 2022 IEEE Conf. on Computer Communications (INFOCOM 2022). London: IEEE, 2022. 660–669. [doi: 10.1109/INFOCOM48880.2022.9796791]
    [53] Zhang L, Zhang CT, Shihada B. Efficient wireless traffic prediction at the edge: A federated meta-learning approach. IEEE Communications Letters, 2022, 26(7): 1573–1577.
    [54] Barlacchi G, De Nadai M, Larcher R, Casella A, Chitic C, Torrisi G, Antonelli F, Vespignani A, Pentland A, Lepri B. A multi-source dataset of urban life in the city of Milan and the province of trentino. Scientific Data, 2015, 2: 150055.
    [55] Telecom Italia. Telecommunications—SMS, Call, Internet—TN, 2015. https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/QLCABU
    [56] Telecom Italia. Telecommunications—SMS, Call, Internet—MI, 2015. https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/EGZHFV
    引证文献
    网友评论
    网友评论
    分享到微博
    发 布
引用本文

柴宝宝,董安明,王桂娟,韩玉冰,李浩,禹继国.基于注意力机制的联邦无线流量预测模型.软件学报,2025,36(2):715-731

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2023-08-09
  • 最后修改日期:2023-10-12
  • 在线发布日期: 2024-07-17
文章二维码
您是第19789626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