Variational Inference for Nonlinear Structural ‎Identification

Document Type : Special Issue Paper

Authors

1 Doctoral Candidate, Lyles School of Civil Engineering, Purdue University, 1040 S. River Road, West LaFayette, IN, 47907, USA

2 Associate Professor, School of Mechanical Engineering, Purdue University, 585 Purdue Mall, West Lafayette, 47907, USA

3 Professor, Lyles School of Civil Engineering and School of Mechanical Engineering, , Purdue University, 585 Purdue Mall, West Lafayette, 47907, USA

Abstract

Research interest in predictive modeling within the structural engineering community has recently been focused on Bayesian inference methods, with particular emphasis on analytical and sampling approaches. In this study, we explore variational inference, a relatively unknown class of Bayesian inference approaches which has potential to realize the computational speed, accuracy, and scalability necessary for structural health monitoring applications. We apply this method to the predictive modeling of a simulated Bouc-Wen system subject to base vibration and compare the performance of this inference approach to that of the unscented Kalman filter. From this investigation, we find that though variational inference is more computationally intensive than the unscented Kalman filter, it exhibits superior performance and flexibility.

Keywords

Main Subjects

[1] Kalman, R.E., A new approach to linear filtering and prediction problems, Journal of Basic Engineering, 82(1), 1960, 35–45, doi: 10.1115/1.3662552.
[2] Schmidt, S.F., The Kalman filter - Its recognition and development for aerospace applications, Journal of Guidance and Control, 4(1), 1981, 4–7, doi: 10.2514/3.19713.
[3] Sorenson, H.W., Kalman filtering techniques, Advances in Control Systems, 3, 1966, 219–292, doi: 10.1016/b978-1-4831-6716-9.50010-2.
[4] Julier, S.J., Uhlmann, J.K., Durrant-Whyte, H.F., A new approach for filtering nonlinear systems, Proceedings of American Control Conference, 1995, 3, 1628–1632, doi: 10.1109/ACC.1995.529783.
[5] Wan, E.A., Van Der Merwe, R., The unscented Kalman filter for nonlinear estimation, IEEE 2000 Adaptive Systems for Signal Processing, Communications, and Control Symposium, 2000, 153–158.
[6] Zhou, L., Wu, S., Yang, J.N., Experimental study of an adaptive extended Kalman filter for structural damage identification, Journal of Infrastructure Systems, 14(1), 2008, 42–51.
[7] Chatzi, E.N., Smyth, A.W., Masri, S.F., Experimental application of on-line parametric identification for nonlinear hysteretic systems with model uncertainty, Structural Safety, 32, 2010, 326–337, doi: 10.1016/j.strusafe.2010.03.008.
[8] Song, W., Dyke, S.J., Real-time dynamic model updating of a hysteretic structural system, Journal of Structural Engineering, 140(3), 2014, 1–14, doi: 10.1061/(ASCE)ST.1943-541X.0000857.
[9] Sarkka, S., Bayesian filtering and smoothing, 1st ed. Cambridge University Press, Cambridge, United Kingdom, 2013.
[10] Olivier, A., Smyth, A.W., On the performance of online parameter estimation algorithms in systems with various identifiability properties, Frontiers in Built Environment, 3(14), 2017, 1–18, doi: 10.3389/fbuil.2017.00014.
[11] Wan, E., van der Merwe, R., Nelson, A.T., Dual estimation and the unscented transformation, Advances in Neural Information Processing Systems, 2000, 666–672.
[12] Lund, A., Dyke, S.J., Song, W., Bilionis, I., Identification of an experimental nonlinear energy sink device using the unscented Kalman filter, Mechanical Systems and Signal Processing, 136, 2020, doi: https://doi.org/10.1016/j.ymssp.2019.106512.
[13] Dertimanis, V.K., Chatzi, E.N., Eftekhar Azam, S., Papadimitriou, C., Input-state-parameter estimation of structural systems from limited output information, Mechanical Systems and Signal Processing, 126, 2019, 711–746, doi: 10.1016/j.ymssp.2019.02.040.
[14] Yuen, K.V., Kuok, S.C., Online updating and uncertainty quantification using nonstationary output-only measurement, Mechanical Systems and Signal Processing, 66, 2016, 62–77, doi: 10.1016/j.ymssp.2015.05.019.
[15] Lei, Y., Xia, D., Erazo, K., Nagarajaiah, S., A novel unscented Kalman filter for recursive state-input-system identification of nonlinear systems, Mechanical Systems and Signal Processing, 127, 2019, 120–135, doi: 10.1016/j.ymssp.2019.03.013.
[16] Solonen, A., Hakkarainen, J., Ilin, A., Abbas, M., Bibov, A., Estimating model error covariance matrix parameters in extended Kalman filtering, Nonlinear Processes in Geophysics, 21(5), 2014, 919–927, doi: 10.5194/npg-21-919-2014.
[17] Erazo, K., Sen, D., Nagarajaiah, S., Sun, L., Vibration-based structural health monitoring under changing environmental conditions using Kalman filtering, Mechanical Systems and Signal Processing, 117, 2019, 1–15, doi: 10.1016/j.ymssp.2018.07.041.
[18] Song, M., Astroza, R., Ebrahimian, H., Moaveni, B., Papadimitriou, C., Adaptive Kalman filters for nonlinear finite element model updating, Mechanical Systems and Signal Processing, 143, 2020, 106837, doi: 10.1016/j.ymssp.2020.106837.
[19] Gordon, N.J., Salmond, D.J., Smith, A.F.M., Novel approach to nonlinear / non-Gaussian Bayesian state estimation, IEE Proceedings (Radar and Signal Processing), 140(2), 1993, 107–113.
[20] Doucet, A., Godsill, S., Andrieu, C., On sequential Monte Carlo sampling methods for Bayesian filtering, Statistics and Computing, 10(3), 2000, 197–208.
[21] Chatzi, E.N., Smyth, A., The unscented Kalman filter and particle filter methods for nonlinear structural system identification with non-collocated heterogeneous sensing, Structural Control and Health Monitoring, 16, 2009, 99–123, doi: 10.1002/stc.
[22] Snyder, C., Bengtsson, T., Bickel, P., Anderson, J., Obstacles to high-dimensional particle filtering, Monthly Weather Review, 136(12), 2008, 4629–4640, doi: 10.1175/2008MWR2529.1.
[23] Chatzi, E., Smyth, A., Particle filter scheme with mutation for the estimation of time-invariant parameters in structural health monitoring applications, Structural Control and Health Monitoring, 20, 2013, 1081–1095, doi: 10.1002/stc.
[24] Olivier, A., Smyth, A.W., Particle filtering and marginalization for parameter identification in structural systems, Structural Control and Health Monitoring, 24, 2017, 1–25, doi: 10.1002/stc.1874.
[25] Azam, S.E., Mariani, S., Dual estimation of partially observed nonlinear structural systems: A particle filter approach, Mechanics Research Communications, 46, 2012, 54–61, doi: 10.1016/j.mechrescom.2012.08.006.
[26] Haug, A.J., Bayesian Estimation and Tracking: A Practical Guide, 1st ed. John Wiley & Sons, Hoboken, New Jersey, 2012.
[27] Blei, D.M., Kucukelbir, A., Mcauliffe, J.D., Variational inference: A review for statisticians, Journal of the American Statistical Association, 112(518), 2017, 859–877, doi: 10.1080/01621459.2017.1285773.
[28] Bishop, C.M., Pattern recognition and machine learning, 1st ed. Springer, Cambridge, UK, 2006.
[29] Saul, L.K., Jaakkola, T., Jordan, M.I., Mean field theory for sigmoid belief networks, Journal of Artificial Intelligence Research, 4, 1996, 61–76.
[30] Saul, L.K., Jordan, M.I., Exploiting Tractable Substructures in Intractable Networks, Advances in Neural Information Processing Systems, 1996, 486–492.
[31] Jaakkola, T.S., Jordan, M.I., A variational approach to Bayesian logistic regression models and their extensions, 6th International Workshop on Artificial Intelligence and Statistics, 1996, 4–15.
[32] Hinton, G.E., Van Camp, D., Keeping Neural Networks Simple by Minimizing the Description Length of the Weights, 6th Annual Conference on Computational Learning Theory, 1993.
[33] Jordan, M.I., Ghahramani, Z., Jaakkola, T.S., Saul, L.K., An introduction to variational methods for graphical models, Machine Learning, 37(2), 1999, 183–233, doi: 10.1007/978-94-011-5014-9_5.
[34] Blei, D.M., Ng, A.Y., Jordan, M.I., Latent Dirichlet allocation, Journal of Machine Learning Research, 3, 2003, 993–1022.
[35] Barber, D., Wiegernick, W., Tractable Variational Structures for Approximating Graphical Models, Advances in Neural Information Processing Systems, 1999, 183–199.
[36] Beal, M.J., Variational Algorithms for Approximate Bayesian Inference, Ph. D. Thesis, University College, London, 2003.
[37] Hoffman, M.D., Blei, D.M., Wang, C., Paisley, J., Stochastic variational inference, Journal of Machine Learning Research, 14(1), 2013, 1303–1347.
[38] Ranganath, R., Gerrish, S., Blei, D.M., Black box variational inference, Artificial Intelligence and Statistics, 2014, 814–822.
[39] Kucukelbir, A., Tran, D., Ranganath, R., Gelman, A., Blei, D.M., Automatic differentiation variational inference, Journal of Machine Learning Research, 18(1), 2017, 430–474.
[40] Gardner, P., Lord, C., Barthorpe, R.J., A unifying framework for probabilistic validation metrics, Journal of Verification, Validation, and Unvertainty Quantification, 2019, doi: 10.1115/1.4045296.
[41] Bamler, R., Zhang, C., Opper, M., Mandt, S., Tightening Bounds for Variational Inference by Revisiting Perturbation Theory, Journal of Statistical Mechanics: Theory and Experiment, (12), 2019, doi: https://doi.org/10.1088/1742-5468/ab43d3.
[42] McInerney, J., Ranganath, R., Blei, D., The population posterior and Bayesian modeling on streams, Advances in Neural Information Processing Systems, 2015, 1153–1161.
[43] Jihan, N., Jayasinghe, M., Perera, S., Streaming stochastic variational Bayes; An improved approach for Bayesian inference with data streams, PeerJ Preprints, 2019, doi: 10.7287/peerj.preprints.27790v2.
[44] Blei, D.M., Lafferty, J.D., A Correlated Topic Model of Science, The Annals of Applied Statistics, 1(1), 2007, 17–35, doi: 10.1214/07-AOAS114.
[45] Cohen, S.B., Smith, N.A., Covariance in unsupervised learning of probabilistic grammars, Journal of Machine Learning Research, 11, 2010, 3017–3051.
[46] Likas, A., Galatsanos, N.P., A variational method for Bayesian blind image deconvolution, International Conference on Image Processing, 2004, 52(8), 2222–2233, doi: 10.1109/ICIP.2003.1246846.
[47] Logsdon, B.A., Hoffman, G.E., Mezey, J.G., A variational Bayes algorithm for fast and accurate multiple locus genome-wide association analysis, BMC Bioninformatics, 11(1), 2010, 58–70.
[48] Raj, A., Stephens, M., Pritchard, J.K., fastSTRUCTURE: variational inference of population structure in large SNP data sets, Genetics, 197(2), 2014, 573–589, doi: 10.1534/genetics.114.164350.
[49] Lund, A., Bilionis, I., Dyke, S.J., Approximate Bayesian approach to rapid structural identification, Proceedings of the 17th World Conference on Earthquake Engineering, 2020.
[50] Wu, M., Smyth, A.W., Application of the unscented Kalman filter for real-time nonlinear structural system identification, Structural Control and Health Monitoring, 14, 2007, 971–990, doi: 10.1002/stc.
[51] Kingma, D.P., Ba, J., Adam: A Method for Stochastic Optimization, International Conference on Learning Representations, 2014, 1–15.
[52] Paszke, A. et al., PyTorch: An Imperative Style, High-Performance Deep Learning Library, Advances in Neural Information Processing Systems, 2019, 8024–8035.
[53] Lund, A., Dyke, S.J., Song, W., Bilionis, I., Global sensitivity analysis for the design of nonlinear identification experiments, Nonlinear Dynamics, 98(1), 2019, 375–394, doi: 10.1007/s11071-019-05199-9.
[54] Saltelli, A. et al., Global Sensitivity Analysis. The Primer, 1st ed. John Wiley & Sons, West Sussex, England, 2008.
[55] Herman, J., Usher, W., SaLib: An open-source Python library for sensitivity analysis, Journal of Open Source Software, 2(9), 2017, doi: doi:10.21105/joss.00097.
[56] Øksendal, B., Stochastic Differential Equations, 5th ed. Springer, Heidelberg, Germany, 2003.
[57] Murphy, K.P., Machine Learning: A Probabilistic Perspective. MIT Press, Cambridge, Massachusetts, 2012.