Abstract: Predicting future heart rate (HR) not only helps in detecting abnormal heart rhythms but also provides timely support for downstream health monitoring services. Existing methods for HR prediction encounter challenges, especially concerning privacy protection and data heterogeneity. To address these challenges, this paper proposes a novel HR prediction framework, PCFedH, which leverages personalized federated learning and prototypical contrastive learning to achieve stable clustering results and more accurate predictions. PCFedH contains two core modules: a prototypical contrastive learning-based federated clustering module, which characterizes data heterogeneity and enhances HR representation to facilitate more effective clustering, and a two-phase soft clustered federated learning module, which enables personalized performance improvements for each local model based on stable clustering results. Experimental results on two real-world datasets demonstrate the superiority of our approach over state-of-the-art methods, achieving an average reduction of 3.1% in the mean squared error across both datasets. Additionally, we conduct comprehensive experiments to empirically validate the effectiveness of the key components in the proposed method. Among these, the personalization component is identified as the most crucial aspect of our design, indicating its substantial impact on overall performance.

面向心率预测的原型聚类联邦学习

[1]Alharbi A, Alosaimi W, Sahal R, et al., 2021. Real-time system prediction for heart rate using deep learning and stream processing platforms. Complexity, 2021:5535734.

[2]Brisimi TS, Chen RD, Mela T, et al., 2018. Federated learning of predictive models from federated electronic health records. Int J Med Inform, 112:59-67.

[3]Cai LX, Chen NY, Cao YZH, et al., 2023. FedCE: personalized federated learning method based on clustering ensembles. Proc 31^st ACM Int Conf on Multimedia, p.1625-1633.

[4]Chen J, Jiang YN, Huang ZX, et al., 2021. Fine-grained detection of driver distraction based on neural architecture search. IEEE Trans Intell Transp Syst, 22(9):5783-5801.

[5]Fang L, Liu XL, Su X, et al., 2020. Bayesian inference federated learning for heart rate prediction. Proc 9^th EAI Int Conf on Wireless Mobile Communication and Healthcare, p.116-130.

[6]Ghosh A, Chung J, Yin D, et al., 2022. An efficient framework for clustered federated learning. IEEE Trans Inform Theory, 68(12):8076-8091.

[7]Jain AK, Dubes RC, 1988. Algorithms for Clustering Data. Prentice-Hall, Inc., Saddle River, USA.

[8]Kingma DP, Ba LJ, 2015. Adam: a method for stochastic optimization. 3^rd Int Conf on Learning Representations, p.1-15.

[9]Li JN, Zhou P, Xiong CM, et al., 2021. Prototypical contrastive learning of unsupervised representations. https://arxiv.org/abs/2005.04966

[10]Li T, Sahu AK, Zaheer M, et al., 2020. Federated optimization in heterogeneous networks. Proc Mach Learn Syst, 2:429-450.

[11]McMahan B, Moore E, Ramage D, et al., 2017. Communication-efficient learning of deep networks from decentralized data. Proc 20^th Int Conf on Artificial Intelligence and Statistics, p.1273-1282.

[12]Mehrang S, Pietila J, Tolonen J, et al., 2017. Human activity recognition using a single optical heart rate monitoring wristband equipped with triaxial accelerometer. European Medical and Biological Engineering Conf, p.587-590.

[13]Mu XT, Shen YL, Cheng K, et al., 2023. FedProc: prototypical contrastive federated learning on non-IID data. Fut Gener Comp Syst, 143:93-104.

[14]Oyeleye M, Chen TH, Titarenko S, et al., 2022. A predictive analysis of heart rates using machine learning techniques. Int J Environ Res Public Health, 19(4):2417.

[15]Panwar M, Gautam A, Biswas D, et al., 2020. PP-net: a deep learning framework for PPG-based blood pressure and heart rate estimation. IEEE Sensors J, 20(17):10000-10011.

[16]Patel M, Lal SKL, Kavanagh D, et al., 2011. Applying neural network analysis on heart rate variability data to assess driver fatigue. Expert Syst Appl, 38(6):7235-7242.

[17]Patidar S, Pachori RB, Acharya UR, 2015. Automated diagnosis of coronary artery disease using tunable-Q wavelet transform applied on heart rate signals. Knowl-Based Syst, 82:1-10.

[18]Qi Z, Meng L, Chen ZT, et al., 2023. Cross-silo prototypical calibration for federated learning with non-IID data. Proc 31^st ACM Int Conf on Multimedia, p.3099-3107.

[19]Rather AM, Agarwal A, Sastry VN, 2015. Recurrent neural network and a hybrid model for prediction of stock returns. Expert Syst Appl, 42(6):3234-3241.

[20]Reiss A, Indlekofer I, Schmidt P, et al., 2019. Deep PPG: large-scale heart rate estimation with convolutional neural networks. Sensors, 19(14):3079.

[21]Ruan YC, Joe-Wong C, 2022. FedSoft: soft clustered federated learning with proximal local updating. Proc 36^th AAAI Conf on Artificial Intelligence, p.8124-8131.

[22]Sattler F, Müller KR, Samek W, 2021. Clustered federated learning: model-agnostic distributed multitask optimization under privacy constraints. IEEE Trans Neur Netw Learn Syst, 32(8):3710-3722.

[23]Staffini A, Svensson T, Chung UI, et al., 2022. Heart rate modeling and prediction using autoregressive models and deep learning. Sensors, 22(1):34.

[24]Taamneh S, Tsiamyrtzis P, Dcosta M, et al., 2017. A multimodal dataset for various forms of distracted driving. Sci Data, 4(1):170110.

[25]Tan AZ, Yu H, Cui LZ, et al., 2023. Towards personalized federated learning. IEEE Trans Neur Netw Learn Syst, 34(12):9587-9603.

[26]Tan Y, Long GD, Ma J, et al., 2022a. Federated learning from pre-trained models: a contrastive learning approach. Proc 36^th Conf on Neural Information Processing System, p.19332-19344.

[27]Tan Y, Long GD, Liu L, et al., 2022b. FedProto: federated prototype learning across heterogeneous clients. Proc 36^th AAAI Conf on Artificial Intelligence, p.8432-8440.

[28]van der Maaten L, Hinton G, 2008. Visualizing data using t-SNE. J Mach Learn Res, 9(86):2579-2605.

[29]Xu MA, Moreno A, Nagesh S, et al., 2022. PulseImpute: a novel benchmark task for pulsative physiological signal imputation. Proc 36^th Conf on Neural Information Processing Systems, p.26874-26888.

[30]Yao HX, Tang XF, Wei H, et al., 2019. Revisiting spatial-temporal similarity: a deep learning framework for traffic prediction. Proc 33^rd AAAI Conf on Artificial Intelligence, p.5668-5675.

[31]Yaqoob MM, Nazir M, Khan MA, et al., 2023. Hybrid classifier-based federated learning in health service providers for cardiovascular disease prediction. Appl Sci, 13(3):1911.

[32]Yu ZT, Shen YM, Shi JG, et al., 2022. PhysFormer: facial video-based physiological measurement with temporal difference transformer. Proc IEEE/CVF Conf on Computer Vision and Pattern Recognition, p.4176-4186.

[33]Yue ZH, Wang YJ, Duan JY, et al., 2022. TS2Vec: towards universal representation of time series. Proc AAAI Conf on Artificial Intelligence, p.8980-8987.

[34]Zhang J, Hua Y, Wang H, et al., 2023. FedALA: adaptive local aggregation for personalized federated learning. Proc AAAI Conf on Artificial Intelligence, p.11237-11244.

[35]Zhu HY, Xu JJ, Liu SQ, et al., 2021. Federated learning on non-IID data: a survey. Neurocomputing, 465:371-390.