HSDC-Net: Privacy-Preserving and Trust-Aware Historical-State Decay Compensation for Secure Distributed Industrial Edge Monitoring
DOI:
https://doi.org/10.4108/eetsis.14417Keywords:
Distributed industrial edge monitoring, historical-state decay, differential privacy, source trust, feature drift, poisoned-anchor suppressionAbstract
INTRODUCTION: Distributed industrial edge monitoring requires long-sequence models to remain reliable under non-stationary operating conditions while preventing sensitive local data from being exposed through cross-node historical-state sharing.
OBJECTIVES: This work addresses secure reuse of historical representations under feature drift, privacy leakage, and poisoned-memory propagation.
METHODS: We propose HSDC-Net, which locally encodes heterogeneous process streams, releases clipped and differentially private latent anchors, estimates cross-node inconsistency and recursive source trust, aligns reliable historical states through low-rank context transport, and restricts correction to a learned nuisance-drift subspace.
RESULTS: Experiments on four industrial datasets show that HSDC-Net consistently improves predictive performance and drift robustness with moderate online latency. It also suppresses malicious-anchor acceptance while maintaining a finite privacy budget with only limited utility degradation relative to the non-private variant.
CONCLUSION: HSDC-Net provides a unified privacy-aware and poisoning-resilient mechanism for robust distributed industrial monitoring without centralizing raw manufacturing streams across heterogeneous monitoring and maintenance tasks.
References
[1] D. Baby, B. Han, S. Zhang, C. Hu, B. Wang, Y.-X. Wang, Adapting to online distribution shifts in deep learning: A black-box approach, in: International Conference on Artificial Intelligence and Statistics, PMLR, 2025, pp. 3853–3861.
[2] F. Regol, L. Schwinn, K. Sprague, M. Coates, T. Markovich, When to retrain a machine learning model, in: International Conference on Machine Learning, PMLR, 2025, pp. 51369–51404.
[3] J.-H. Lee, Bayesian weight enhancement with steady-state adaptation for test-time adaptation in dynamic environments, in: International Conference on Machine Learning, PMLR, 2025, pp. 32889–32906.
[4] J. Liu, N. Seedat, P. Cui, M. van der Schaar, Going beyond static: Understanding shifts with time-series attribution, in: The Thirteenth International Conference on Learning Representations, 2025. URL https://openreview.net/forum?id=XQlccqJpCC
[5] L. Lin, X. Zhang, Q. Zhang, S. Wang, K. Xia, Ortcl: Towards continual learning of time series foundation models on streaming data via orthogonal rotation, in: Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 40, 2026, pp. 23541–23549.
[6] D. Liang, J. Chen, X. Wang, Y. Wang, S. Li, Deepboots: Dual-stream residual boosting for drift-resilient time-series forecasting, in: Proceedings of the AAAI Confer-ence on Artificial Intelligence, Vol. 40, 2026, pp. 23391–23400.
[7] J. Chapelin, A. Voisin, B. Rose, B. Iung, L. Steck, L. Chaves, M. Lauer, O. Jotz, Data-driven drift detection and diagnosis framework for predictive maintenance of heterogeneous production processes: Application to a multiple tapping process, Engineering Applications of Artificial Intelligence 139 (2025) 109552.
[8] M. A. Benatia, M. Hafsi, S. B. Ayed, A continual learning approach for failure prediction under non-stationary conditions: Application to condition monitoring data streams, Computers & Industrial Engineering 204 (2025) 111049.
[9] D. M. V. Sato, E. Ruschel, E. E. Scalabrin, E. d. F. R. Loures, E. A. P. Santos, Interactive process drift detection (ipdd) for condition-based maintenance using process mining, Advanced Engineering Informatics 66 (2025) 103397.
[10] S. Li, B. Zhou, J. Shang, X. Chen, J. Yu, Review of recent applications and future perspectives on process monitoring approaches in industrial processes, Journal of Manufacturing Systems 82 (2025) 509–530.
[11] T. Özel, Deep learning-based applications in metal additive manufacturing processes: Challenges and opportunities–a review, International Journal of Lightweight Materials and Manufacture 8 (4) (2025) 453–468.
[12] S. Oueslati, M. Ritou, E. Paquet, F. Belkadi, P. Le Bot, Prediction of process instability by waam in-process monitoring and ctwd drift estimation, Additive Manu-facturing (2026) 105127.
[13] Y. Mao, J. Yan, X. Lin, C. K. Chua, K. Zhu, A physics-guided multi-source data fusion model for lpbf process monitoring: From data generation to melting state recognition, Additive Manufacturing (2026) 105200.
[14] A. Caggiano, G. Mattera, Y. Zhang, R. Teti, Heat input control and deep learning-based indirect measure of process and deposition stability in wire arc additive manufacturing, CIRP Annals (2026).
[15] S. Oueslati, E. Paquet, F. Belkadi, M. Ritou, P. Le Bot, New monitoring criteria for instability detection by machine learning in additive manufacturing: Application to wire arc additive manufacturing process, IFAC-PapersOnLine 59 (10) (2025) 2250–2255.
[16] J. Xie, M. Safdar, A. M. Romascanu, Y. Lu, H. Ko, Z. Yang, Y. F. Zhao, Towards reproducible machine learning-based process monitoring and quality predic-tion research for additive manufacturing, in: Interna-tional design engineering technical conferences and computers and information in engineering conference, Vol. 88346, American Society of Mechanical Engineers, 2024, p. V02AT02A033.
[17] Q. Liu, R. Yin, X. Guo, W. Wang, Z. Gao, M. Jia, Y. Liu, Locally spatiotemporal soft sensor for key indicator prediction in cement production process, Chemical Engineering Science 307 (2025) 121386.
[18] J. Li, X. Lin, G. S. Hong, K. Zhu, Uncovering nonequilibrium dynamics in machining via mechanism-assisted machine learning, npj Advanced Manufacturing 2 (1) (2025) 33.
[19] G. Mattera, R. Mattera, L. Nele, A forecast-assisted approach to remaining useful life prediction: A pre-dictive maintenance case study in hybrid al/cfrp stack drilling, International Journal of Production Research 64 (9) (2026) 3400–3422.
[20] S. Hochreiter, J. Schmidhuber, Long short-term memory, Neural Computation 9 (8) (1997) 1735–1780. doi:10. 1162/neco.1997.9.8.1735.
[21] H. Zhou, S. Zhang, J. Peng, S. Zhang, J. Li, H. Xiong, W. Zhang, Informer: Beyond efficient transformer for long sequence time-series forecasting, in: Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 35, 2021, pp. 11106–11115. doi:10.1609/aaai.v35i12. 17325.
[22] Y. Nie, N. H. Nguyen, P. Sinthong, J. Kalagnanam, A time series is worth 64 words: Long-term forecasting with transformers, in: The Eleventh International Conference on Learning Representations, 2023. URL https://openreview.net/forum?id= Jbdc0vTOcol
[23] F. M. Abushaqra, H. Xue, Y. Ren, F. D. Salim, ODEStream: A buffer-free online learning framework with ODE-based adaptor for streaming time series forecasting, Transactions on Machine Learning Research (2025). URL https://openreview.net/forum?id=TWOTKhwU5n
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Jianxia Liu, Zicheng Zheng
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
This is an open access article distributed under the terms of the CC BY-NC-SA 4.0, which permits copying, redistributing, remixing, transformation, and building upon the material in any medium so long as the original work is properly cited.
