Graph-based Active Learning for Dynamic Industrial Systems with Temporal Evolution
DOI:
https://doi.org/10.63282/3050-9246.IJETCSIT-V1I1P109Keywords:
Graph Neural Networks, Active Learning, Federated Learning, Explainability, Anomaly Detection, Predictive Maintenance, Temporal Graphs, Industrial IoTAbstract
In the era of Industry 4.0, dynamic industrial systems characterized by temporal evolution present significant challenges for anomaly detection and predictive maintenance. This paper proposes a novel graph-based active learning frame- work that integrates time-evolving node features to enhance model performance in such environments. We introduce a trust metric-based federated learning architecture to ensure integrity and accountability across distributed data silos, addressing privacy and collaboration issues in industrial IoT. Additionally, we develop a quantification and optimization framework for the trade-off between model explainability and performance, enabling practitioners to balance these often conflicting objectives. Through extensive experiments on benchmark datasets and simulated industrial scenarios, we demonstrate the superiority of our approach in terms of accuracy, efficiency, and interpretability. Our results show improvements of up to 15% in anomaly detection precision while maintaining high explainability scores. This work contributes to the advancement of machine learning applications in dynamic industrial systems, providing a comprehensive solution that is both practical and theoretically grounded
Downloads
References
[1] Settles, B. (2010). Active Learning Literature Survey. University of Wisconsin–Madison. Burr Settles
[2] Zhu, X., Ghahramani, Z., & Lafferty, J. (2003). Semi-Supervised Learning Using Gaussian Fields and Harmonic Functions. In Proceedings of the 20th International Conference on Machine Learning (ICML), 912–919. DBLP
[3] Zhu, X. (2005). Semi-Supervised Learning Literature Survey (Technical Report). University of Wisconsin–Madison. UW Computer Sciences
[4] Zhu, X., & Ghahramani, Z. (2002). Learning from Labeled and Unlabeled Data with Label Propagation. (Tech. Report / conference preprint). Semantic Scholar
[5] Sen, P., Namata, G., Bilgic, M., Getoor, L., Gallagher, B., & Eliassi-Rad, T. (2008). Collective Classification in Network Data. AI Magazine, 29(3), 93–106. ojs.aaai.org
[6] Kipf, T. N., & Welling, M. (2016). Semi-Supervised Classification with Graph Convolutional Networks. arXiv:1609.02907. arXiv
[7] Hamilton, W. L., Ying, R., & Leskovec, J. (2017). Inductive Representation Learning on Large Graphs (GraphSAGE). In Proceedings of the 31st Conference on Neural Information Processing Systems (NeurIPS) / arXiv:1706.02216. arXiv
[8] Bilgic, M., & Getoor, L. (2010). Active Learning for Networked Data. In Proceedings of the 27th International Conference on Machine Learning (ICML 2010). icml.cc
[9] Macskassy, S. A. (2009). Using Graph-Based Metrics with Empirical Risk Minimization to Speed Up Active Learning on Networked Data. In Proceedings of the 15th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD 2009). ResearchGate
[10] Guillory, A., & Bilmes, J. (2011). Active Semi-Supervised Learning Using Submodular Functions. In Proceedings of the Conference on Uncertainty in Artificial Intelligence (UAI 2011). ACM Digital Library
[11] Gu, Q., & Han, J. (2012). Towards Active Learning on Graphs: An Error Bound Minimization Approach. In Proceedings of ICDM 2012. hanj.cs.illinois.edu
[12] Cesa-Bianchi, N., Gentile, C., & Zappella, G. (2013). Active Learning on Trees and Graphs. (COLT / arXiv / conference version). arXiv
[13] Ma, Y., Garnett, R., & Schneider, J. (2013). Σ-Optimality for Active Learning on Gaussian Random Fields. (Workshop / conference paper on active learning and GRF). CMU School of Computer Science
[14] Guillory, A., & Bilmes, J. (2010). Interactive Submodular Set Cover. In Proceedings of ICML 2010 / NeurIPS related work (online submodular/set-cover and active learning). papers.nips.cc
[15] Rattigan, M., Maier, M., & Jensen, D. (2007). Exploiting Network Structure for Active Inference in Collective Classification. In Proceedings of the 7th ICDM Workshops (ICDMW 2007).
[16] Cohn, D. A., Ghahramani, Z., & Jordan, M. I. (1996). Active Learning with Statistical Models. arXiv
[17] Tong, S., & Koller, D. (2001). Support Vector Machine Active Learning with Applications to Text Classification. Journal of Machine Learning Research. jmlr.org
[18] Zhou, D., Bousquet, O., Lal, T. N., Weston, J., & Schölkopf, B. (2004). Learning with Local and Global Consistency. NeurIPS. papers.nips.cc
[19] Talukdar, P. P., & Pereira, F. (2010). Experiments in Graph-Based Semi-Supervised Learning. EMNLP. ACL Anthology
[20] Talukdar, P. P., & Crammer, K. (2009). New Regularized Algorithms for Transductive Learning. ECML-PKDD. repository.upenn.edu
[21] Perozzi, B., Al-Rfou, R., & Skiena, S. (2014). DeepWalk: Online Learning of Social Representations. KDD. arXiv
[22] Grover, A., & Leskovec, J. (2016). node2vec: Scalable Feature Learning for Networks. KDD. arXiv
[23] Tang, J., Qu, M., Wang, M., Zhang, M., Yan, J., & Mei, Q. (2015). LINE: Large-scale Information Network Embedding. WWW. ACM Digital Library
[24] Goyal, P., Kamra, N., He, X., & Liu, Y. (2018). DynGEM: Deep Embedding Method for Dynamic Graphs. arXiv preprint. arXiv
