Robustness and Generalization Performance of Deep Learning Models on Cyber-Physical Systems: A Comparative Study

Deep learning (DL) models have seen increased attention for time series forecasting, yet the application on cyber-physical systems (CPS) is hindered by the lacking robustness of these methods. Thus, this study evaluates the robustness and generalization performance of DL architectures on multivariate time series data from CPS. Our investigation focuses on the models' ability to handle a range of perturbations, such as sensor faults and noise, and assesses their impact on overall performance. Furthermore, we test the generalization and transfer learning capabilities of these models by exposing them to out-of-distribution (OOD) samples. These include deviations from standard system operations, while the core dynamics of the underlying physical system are preserved. Additionally, we test how well the models respond to several data augmentation techniques, including added noise and time warping. Our experimental framework utilizes a simulated three-tank system, proposed as a novel benchmark for evaluating the robustness and generalization performance of DL algorithms in CPS data contexts. The findings reveal that certain DL model architectures and training techniques exhibit superior effectiveness in handling OOD samples and various perturbations. These insights have significant implications for the development of DL models that deliver reliable and robust performance in real-world CPS applications.

[1]  Roman Kern,et al.  Constructing robust health indicators from complex engineered systems via anticausal learning , 2022, Eng. Appl. Artif. Intell..

[2]  Meng Jiang,et al.  Graph Data Augmentation for Graph Machine Learning: A Survey , 2022, IEEE Data Eng. Bull..

[3]  Yuki M. Asano,et al.  CITRIS: Causal Identifiability from Temporal Intervened Sequences , 2022, ICML.

[4]  O. Niggemann,et al.  A Research Agenda for AI Planning in the Field of Flexible Production Systems , 2021, 2022 IEEE 5th International Conference on Industrial Cyber-Physical Systems (ICPS).

[5]  Hritik Bansal,et al.  Systematic Generalization in Neural Networks-based Multivariate Time Series Forecasting Models , 2021, 2021 International Joint Conference on Neural Networks (IJCNN).

[6]  Christian Reuter,et al.  A Survey on Data Augmentation for Text Classification , 2021, ACM Comput. Surv..

[7]  Geoffrey E. Hinton,et al.  Deep learning for AI , 2021, Commun. ACM.

[8]  Oliver Niggemann,et al.  Generating Artificial Sensor Data for the Comparison of Unsupervised Machine Learning Methods , 2021, Sensors.

[9]  Antonella Longo,et al.  Leveraging Data Augmentation for Service QoS Prediction in Cyber-physical Systems , 2021, ACM Trans. Internet Techn..

[10]  Nan Rosemary Ke,et al.  Coordination Among Neural Modules Through a Shared Global Workspace , 2021, ICLR.

[11]  Yoshua Bengio,et al.  Towards Causal Representation Learning , 2021, ArXiv.

[12]  Anuradha Bhamidipaty,et al.  A Transformer-based Framework for Multivariate Time Series Representation Learning , 2020, KDD.

[13]  N. Zabaras,et al.  Transformers for Modeling Physical Systems , 2020, Neural Networks.

[14]  Mark Chen,et al.  Language Models are Few-Shot Learners , 2020, NeurIPS.

[15]  Chao Meng,et al.  A Time Convolutional Network Based Outlier Detection for Multidimensional Time Series in Cyber-Physical-Social Systems , 2020, IEEE Access.

[16]  D. Yao,et al.  Deep Learning-Based Anomaly Detection in Cyber-Physical Systems: Progress and Opportunities , 2020, ACM Comput. Surv..

[17]  Stavros Tripakis,et al.  Metrics and methods for robustness evaluation of neural networks with generative models , 2020, Machine Learning.

[18]  Xiaomin Song,et al.  Time Series Data Augmentation for Deep Learning: A Survey , 2020, IJCAI.

[19]  Zhen Xiao,et al.  Adversarial Attacks and Defenses on Cyber–Physical Systems: A Survey , 2020, IEEE Internet of Things Journal.

[20]  Thomas G. Dietterich What is machine learning? , 2015, Archives of Disease in Childhood.

[21]  Sergey Levine,et al.  Recurrent Independent Mechanisms , 2019, ICLR.

[22]  Taghi M. Khoshgoftaar,et al.  A survey on Image Data Augmentation for Deep Learning , 2019, Journal of Big Data.

[23]  Wenhu Chen,et al.  Enhancing the Locality and Breaking the Memory Bottleneck of Transformer on Time Series Forecasting , 2019, NeurIPS.

[24]  IEEE Access , 2018, IEEE Consumer Electronics Magazine.

[25]  Aleksander Madry,et al.  Robustness May Be at Odds with Accuracy , 2018, ICLR.

[26]  Lukasz Kaiser,et al.  Attention is All you Need , 2017, NIPS.

[27]  Gregory D. Hager,et al.  Temporal Convolutional Networks for Action Segmentation and Detection , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[28]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[29]  Yoshua Bengio,et al.  Learning Phrase Representations using RNN Encoder–Decoder for Statistical Machine Translation , 2014, EMNLP.

[30]  Pascal Vincent,et al.  Representation Learning: A Review and New Perspectives , 2012, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[31]  E. F. Vogel,et al.  A plant-wide industrial process control problem , 1993 .

[32]  H. Garnier,et al.  Unsupervised Remaining Useful Life Prediction through Long Range Health Index Estimation based on Encoders-Decoders , 2022, IFAC-PapersOnLine.

[33]  Henrik S. Steude,et al.  Machine Learning for Cyber-Physical Systems , 2022, Digital Transformation.

[34]  N. Lu,et al.  A Data-Driven Approach for Assessing Aero-Engine Health Status , 2022, IFAC-PapersOnLine.

[35]  Yuki M. Asano,et al.  iCITRIS: Causal Representation Learning for Instantaneous Temporal Effects , 2022, ArXiv.

[36]  Shinichi Nakasuka,et al.  Robustness of AI-based prognostic and systems health management , 2021, Annu. Rev. Control..

[37]  Jay Lee,et al.  Prognostics and health management design for rotary machinery systems—Reviews, methodology and applications , 2014 .

[38]  V. Rico-Ramirez,et al.  Computers & Chemical Engineering , 2013 .

[39]  Léon Bottou,et al.  Stochastic Gradient Descent Tricks , 2012, Neural Networks: Tricks of the Trade.

[40]  Marco Wiering,et al.  2011 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS (IJCNN) , 2011, IJCNN 2011.

[41]  I. Campbell,et al.  Volume 30 , 2002 .