Skip to main content
Top

Hint

Swipe to navigate through the chapters of this book

2020 | OriginalPaper | Chapter

Learning Stochastic Dynamical Systems via Bridge Sampling

Authors : Harish S. Bhat, Shagun Rawat

Published in: Advanced Analytics and Learning on Temporal Data

Publisher: Springer International Publishing

share
SHARE

Abstract

We develop algorithms to automate discovery of stochastic dynamical system models from noisy, vector-valued time series. By discovery, we mean learning both a nonlinear drift vector field and a diagonal diffusion matrix for an Itô stochastic differential equation in \(\mathbb {R}^d\). We parameterize the vector field using tensor products of Hermite polynomials, enabling the model to capture highly nonlinear and/or coupled dynamics. We solve the resulting estimation problem using expectation maximization (EM). This involves two steps. We augment the data via diffusion bridge sampling, with the goal of producing time series observed at a higher frequency than the original data. With this augmented data, the resulting expected log likelihood maximization problem reduces to a least squares problem. We provide an open-source implementation of this algorithm. Through experiments on systems with dimensions one through eight, we show that this EM approach enables accurate estimation for multiple time series with possibly irregular observation times. We study how the EM method performs as a function of the amount of data augmentation, as well as the volume and noisiness of the data.
Literature
1.
go back to reference Archambeau, C., Opper, M., Shen, Y., Cornford, D., Shawe-Taylor, J.S.: Variational inference for diffusion processes. In: Advances in Neural Information Processing Systems, pp. 17–24 (2008) Archambeau, C., Opper, M., Shen, Y., Cornford, D., Shawe-Taylor, J.S.: Variational inference for diffusion processes. In: Advances in Neural Information Processing Systems, pp. 17–24 (2008)
2.
go back to reference Batz, P., Ruttor, A., Opper, M.: Variational estimation of the drift for stochastic differential equations from the empirical density. J. Stat. Mech: Theory Exp. 2016(8), 083404 (2016) MathSciNetCrossRef Batz, P., Ruttor, A., Opper, M.: Variational estimation of the drift for stochastic differential equations from the empirical density. J. Stat. Mech: Theory Exp. 2016(8), 083404 (2016) MathSciNetCrossRef
3.
go back to reference Batz, P., Ruttor, A., Opper, M.: Approximate Bayes learning of stochastic differential equations. Phys. Rev. E 98, 022109 (2018) MathSciNetCrossRef Batz, P., Ruttor, A., Opper, M.: Approximate Bayes learning of stochastic differential equations. Phys. Rev. E 98, 022109 (2018) MathSciNetCrossRef
4.
go back to reference Bhat, H.S., Madushani, R.W.M.A.: Nonparametric adjoint-based inference for stochastic differential equations. In: 2016 IEEE International Conference on Data Science and Advanced Analytics (DSAA), pp. 798–807 (2016) Bhat, H.S., Madushani, R.W.M.A.: Nonparametric adjoint-based inference for stochastic differential equations. In: 2016 IEEE International Conference on Data Science and Advanced Analytics (DSAA), pp. 798–807 (2016)
5.
go back to reference Bhattacharya, R.N., Waymire, E.C.: Stochastic Processes with Applications. SIAM (2009) Bhattacharya, R.N., Waymire, E.C.: Stochastic Processes with Applications. SIAM (2009)
6.
go back to reference Brunton, S.L., Proctor, J.L., Kutz, J.N.: Discovering governing equations from data by sparse identification of nonlinear dynamical systems. Proc. Nat. Acad. Sci. 113(15), 3932–3937 (2016) MathSciNetCrossRef Brunton, S.L., Proctor, J.L., Kutz, J.N.: Discovering governing equations from data by sparse identification of nonlinear dynamical systems. Proc. Nat. Acad. Sci. 113(15), 3932–3937 (2016) MathSciNetCrossRef
7.
go back to reference Chen, S., Shojaie, A., Witten, D.M.: Network reconstruction from high-dimensional ordinary differential equations. J. Am. Stat. Assoc. 112(520), 1697–1707 (2017) MathSciNetCrossRef Chen, S., Shojaie, A., Witten, D.M.: Network reconstruction from high-dimensional ordinary differential equations. J. Am. Stat. Assoc. 112(520), 1697–1707 (2017) MathSciNetCrossRef
8.
go back to reference Ghahramani, Z., Roweis, S.T.: Learning nonlinear dynamical systems using an EM algorithm. In: Advances in Neural Information Processing Systems (NIPS), pp. 431–437 (1999) Ghahramani, Z., Roweis, S.T.: Learning nonlinear dynamical systems using an EM algorithm. In: Advances in Neural Information Processing Systems (NIPS), pp. 431–437 (1999)
10.
go back to reference Mangan, N.M., Brunton, S.L., Proctor, J.L., Kutz, J.N.: Inferring biological networks by sparse identification of nonlinear dynamics. IEEE Trans. Mol. Biol. Multi-Scale Commun. 2(1), 52–63 (2016) CrossRef Mangan, N.M., Brunton, S.L., Proctor, J.L., Kutz, J.N.: Inferring biological networks by sparse identification of nonlinear dynamics. IEEE Trans. Mol. Biol. Multi-Scale Commun. 2(1), 52–63 (2016) CrossRef
11.
go back to reference Mangan, N.M., Kutz, J.N., Brunton, S.L., Proctor, J.L.: Model selection for dynamical systems via sparse regression and information criteria. Proc. R. Soc. A 473(2204), 20170009 (2017) MathSciNetCrossRef Mangan, N.M., Kutz, J.N., Brunton, S.L., Proctor, J.L.: Model selection for dynamical systems via sparse regression and information criteria. Proc. R. Soc. A 473(2204), 20170009 (2017) MathSciNetCrossRef
12.
go back to reference van der Meulen, F., Schauer, M., van Waaij, J.: Adaptive nonparametric drift estimation for diffusion processes using Faber-Schauder expansions. Statistical Inference for Stochastic Processes, pp. 1–26 (2017) van der Meulen, F., Schauer, M., van Waaij, J.: Adaptive nonparametric drift estimation for diffusion processes using Faber-Schauder expansions. Statistical Inference for Stochastic Processes, pp. 1–26 (2017)
13.
go back to reference van der Meulen, F., Schauer, M., van Zanten, H.: Reversible jump MCMC for nonparametric drift estimation for diffusion processes. Comput. Stat. Data Anal. 71, 615–632 (2014) MathSciNetCrossRef van der Meulen, F., Schauer, M., van Zanten, H.: Reversible jump MCMC for nonparametric drift estimation for diffusion processes. Comput. Stat. Data Anal. 71, 615–632 (2014) MathSciNetCrossRef
14.
15.
go back to reference Nicolau, J.: Nonparametric estimation of second-order stochastic differential equations. Econ. Theory 23(05), 880 (2007) MathSciNetCrossRef Nicolau, J.: Nonparametric estimation of second-order stochastic differential equations. Econ. Theory 23(05), 880 (2007) MathSciNetCrossRef
16.
go back to reference Papaspiliopoulos, O., Roberts, G.O.: Importance sampling techniques for estimation of diffusion models. Stat. Methods Stoch. Differ. Equ. 124, 311–340 (2012) MathSciNetMATH Papaspiliopoulos, O., Roberts, G.O.: Importance sampling techniques for estimation of diffusion models. Stat. Methods Stoch. Differ. Equ. 124, 311–340 (2012) MathSciNetMATH
17.
go back to reference Papaspiliopoulos, O., Roberts, G.O., Stramer, O.: Data augmentation for diffusions. J. Comput. Graph. Stat. 22(3), 665–688 (2013) MathSciNetCrossRef Papaspiliopoulos, O., Roberts, G.O., Stramer, O.: Data augmentation for diffusions. J. Comput. Graph. Stat. 22(3), 665–688 (2013) MathSciNetCrossRef
18.
go back to reference Quade, M., Abel, M., Kutz, J.N., Brunton, S.L.: Sparse identification of nonlinear dynamics for rapid model recovery. Chaos 28(6), 063116 (2018) MathSciNetCrossRef Quade, M., Abel, M., Kutz, J.N., Brunton, S.L.: Sparse identification of nonlinear dynamics for rapid model recovery. Chaos 28(6), 063116 (2018) MathSciNetCrossRef
19.
go back to reference Raissi, M., Karniadakis, G.E.: Hidden physics models: machine learning of nonlinear partial differential equations. J. Comput. Phys. 357, 125–141 (2018) MathSciNetCrossRef Raissi, M., Karniadakis, G.E.: Hidden physics models: machine learning of nonlinear partial differential equations. J. Comput. Phys. 357, 125–141 (2018) MathSciNetCrossRef
20.
go back to reference Raissi, M., Perdikaris, P., Karniadakis, G.E.: Machine learning of linear differential equations using Gaussian processes. J. Comput. Phys. 348, 683–693 (2017) MathSciNetCrossRef Raissi, M., Perdikaris, P., Karniadakis, G.E.: Machine learning of linear differential equations using Gaussian processes. J. Comput. Phys. 348, 683–693 (2017) MathSciNetCrossRef
21.
go back to reference Rawat, S.: Learning governing equations for stochastic dynamical systems. Ph.D. thesis, University of California, Merced (2018). Advisor: H.S. Bhat Rawat, S.: Learning governing equations for stochastic dynamical systems. Ph.D. thesis, University of California, Merced (2018). Advisor: H.S. Bhat
22.
go back to reference Raziperchikolaei, R., Bhat, H.: A block coordinate descent proximal method for simultaneous filtering and parameter estimation. In: Chaudhuri, K., Salakhutdinov, R. (eds.) Proceedings of the 36th International Conference on Machine Learning. Proceedings of Machine Learning Research, Long Beach, California, USA, 09–15 June 2019, vol. 97, pp. 5380–5388. PMLR (2019) Raziperchikolaei, R., Bhat, H.: A block coordinate descent proximal method for simultaneous filtering and parameter estimation. In: Chaudhuri, K., Salakhutdinov, R. (eds.) Proceedings of the 36th International Conference on Machine Learning. Proceedings of Machine Learning Research, Long Beach, California, USA, 09–15 June 2019, vol. 97, pp. 5380–5388. PMLR (2019)
23.
go back to reference Roberts, G.O., Stramer, O.: On inference for partially observed nonlinear diffusion models using the Metropolis-Hastings algorithm. Biometrika 88(3), 603–621 (2001) MathSciNetCrossRef Roberts, G.O., Stramer, O.: On inference for partially observed nonlinear diffusion models using the Metropolis-Hastings algorithm. Biometrika 88(3), 603–621 (2001) MathSciNetCrossRef
24.
go back to reference Rudy, S.H., Brunton, S.L., Proctor, J.L., Kutz, J.N.: Data-driven discovery of partial differential equations. Sci. Adv. 3(4), e1602614 (2017) CrossRef Rudy, S.H., Brunton, S.L., Proctor, J.L., Kutz, J.N.: Data-driven discovery of partial differential equations. Sci. Adv. 3(4), e1602614 (2017) CrossRef
25.
go back to reference Ruttor, A., Batz, P., Opper, M.: Approximate Gaussian process inference for the drift function in stochastic differential equations. In: Advances in Neural Information Processing Systems, pp. 2040–2048 (2013) Ruttor, A., Batz, P., Opper, M.: Approximate Gaussian process inference for the drift function in stochastic differential equations. In: Advances in Neural Information Processing Systems, pp. 2040–2048 (2013)
26.
go back to reference Schaeffer, H., Caflisch, R., Hauck, C.D., Osher, S.: Sparse dynamics for partial differential equations. Proc. Nat. Acad. Sci. 110(17), 6634–6639 (2013) MathSciNetCrossRef Schaeffer, H., Caflisch, R., Hauck, C.D., Osher, S.: Sparse dynamics for partial differential equations. Proc. Nat. Acad. Sci. 110(17), 6634–6639 (2013) MathSciNetCrossRef
27.
go back to reference Schaeffer, H.: Learning partial differential equations via data discovery and sparse optimization. Proc. R. Soc. A: Math., Phys. Eng. Sci. 473(2197), 20160446 (2017) MathSciNetCrossRef Schaeffer, H.: Learning partial differential equations via data discovery and sparse optimization. Proc. R. Soc. A: Math., Phys. Eng. Sci. 473(2197), 20160446 (2017) MathSciNetCrossRef
28.
go back to reference Schaeffer, H., Tran, G., Ward, R.: Extracting sparse high-dimensional dynamics from limited data. SIAM J. Appl. Math. 78(6), 3279–3295 (2018) MathSciNetCrossRef Schaeffer, H., Tran, G., Ward, R.: Extracting sparse high-dimensional dynamics from limited data. SIAM J. Appl. Math. 78(6), 3279–3295 (2018) MathSciNetCrossRef
29.
go back to reference Schauer, M., van der Meulen, F., van Zanten, H.: Guided proposals for simulating multi-dimensional diffusion bridges. Bernoulli 23(4A), 2917–2950 (2017) MathSciNetCrossRef Schauer, M., van der Meulen, F., van Zanten, H.: Guided proposals for simulating multi-dimensional diffusion bridges. Bernoulli 23(4A), 2917–2950 (2017) MathSciNetCrossRef
30.
go back to reference Schön, T.B., Svensson, A., Murray, L., Lindsten, F.: Probabilistic learning of nonlinear dynamical systems using sequential Monte Carlo. Mech. Syst. Signal Process. 104, 866–883 (2018) CrossRef Schön, T.B., Svensson, A., Murray, L., Lindsten, F.: Probabilistic learning of nonlinear dynamical systems using sequential Monte Carlo. Mech. Syst. Signal Process. 104, 866–883 (2018) CrossRef
32.
go back to reference Tran, G., Ward, R.: Exact recovery of chaotic systems from highly corrupted data. Multiscale Model. Simul. 15(3), 1108–1129 (2017) MathSciNetCrossRef Tran, G., Ward, R.: Exact recovery of chaotic systems from highly corrupted data. Multiscale Model. Simul. 15(3), 1108–1129 (2017) MathSciNetCrossRef
33.
go back to reference Verzelen, N., Tao, W., Müller, H.G.: others: Inferring stochastic dynamics from functional data. Biometrika 99(3), 533–550 (2012) MathSciNetCrossRef Verzelen, N., Tao, W., Müller, H.G.: others: Inferring stochastic dynamics from functional data. Biometrika 99(3), 533–550 (2012) MathSciNetCrossRef
34.
go back to reference Vrettas, M.D., Opper, M., Cornford, D.: Variational mean-field algorithm for efficient inference in large systems of stochastic differential equations. Phys. Rev. E 91(1), 012148 (2015) MathSciNetCrossRef Vrettas, M.D., Opper, M., Cornford, D.: Variational mean-field algorithm for efficient inference in large systems of stochastic differential equations. Phys. Rev. E 91(1), 012148 (2015) MathSciNetCrossRef
Metadata
Title
Learning Stochastic Dynamical Systems via Bridge Sampling
Authors
Harish S. Bhat
Shagun Rawat
Copyright Year
2020
DOI
https://doi.org/10.1007/978-3-030-39098-3_14

Premium Partner