Abstract
The concept of quasi-potential plays a central role in understanding the mechanisms of rare events and characterizing the statistics of transition behaviors in stochastic dynamics. Despite its significance, the computation of quasi-potential is a challenging problem with limited existing techniques. In this paper, we devise a machine learning method to compute the quasi-potential of stochastic dynamical systems. More specifically, we first derive the Hamilton-Jacobi equation satisfied by quasi-potential via WKB approximation for two important classes of stochastic processes: diffusion processes and continuous-time jump Markov processes. Then we design a neural network with automatic differentiation technique to compute the quasi-potential based on the Hamilton-Jacobi equation. Two prototypical examples are presented to demonstrate the efficacy and accuracy of our method for stochastic differential equations with additive and multiplicative Gaussian noise and continuous-time jump Markov processes. This approach provides an effective tool in exploring the internal mechanisms of rare events triggered by random perturbations in various scientific fields.
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs11071-022-07536-x/MediaObjects/11071_2022_7536_Fig1_HTML.png)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs11071-022-07536-x/MediaObjects/11071_2022_7536_Fig2_HTML.png)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs11071-022-07536-x/MediaObjects/11071_2022_7536_Fig3_HTML.png)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs11071-022-07536-x/MediaObjects/11071_2022_7536_Fig4_HTML.png)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs11071-022-07536-x/MediaObjects/11071_2022_7536_Fig5_HTML.png)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs11071-022-07536-x/MediaObjects/11071_2022_7536_Fig6_HTML.png)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs11071-022-07536-x/MediaObjects/11071_2022_7536_Fig7_HTML.png)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs11071-022-07536-x/MediaObjects/11071_2022_7536_Fig8_HTML.png)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs11071-022-07536-x/MediaObjects/11071_2022_7536_Fig9_HTML.png)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs11071-022-07536-x/MediaObjects/11071_2022_7536_Fig10_HTML.png)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs11071-022-07536-x/MediaObjects/11071_2022_7536_Fig11_HTML.png)
Similar content being viewed by others
Data Availability
Data sharing is not applicable to this article as no datasets were generated or analysed during the current study.
References
Allen, R.J., Warren, P.B., Ten Wolde, P.R.: Sampling rare switching events in biochemical networks. Phys. Rev. Lett. 94(1), 018104 (2005)
Beri, S., Mannella, R., Luchinsky, D.G., Silchenko, A., McClintock, P.V.: Solution of the boundary value problem for optimal escape in continuous stochastic systems and maps. Phys. Rev. E 72(3), 036131 (2005)
Boninsegna, L., Nüske, F., Clementi, C.: Sparse learning of stochastic dynamical equations. J Chem. Phys. 148(24), 241723 (2018)
Bressloff, P.C., Newby, J.M.: Path integrals and large deviations in stochastic hybrid systems. Phys. Rev. E 89(4), 042701 (2014)
Cameron, M.: Finding the quasipotential for nongradient sdes. Physica D: Nonlinear Phenomena 241(18), 1532–1550 (2012)
Chen, X., Wu, F., Duan, J., Kurths, J., Li, X.: Most probable dynamics of a genetic regulatory network under stable lévy noise. Appl. Math. Comput. 348, 425–436 (2019)
Chen, X., Yang, L., Duan, J., Karniadakis, G.E.: Solving inverse stochastic problems from discrete particle observations using the fokker-planck equation and physics-informed neural networks. SIAM J. Sci. Comput. 43(3), B811–B830 (2021)
Chen, Z., Liu, X.: Singularities of fluctuational paths for an overdamped two-well system driven by white noise. Phys A: Stat. Mech. Appl. 469, 206–215 (2017)
Chen, Z., Zhu, J., Liu, X.: Non-differentiability of quasi-potential and non-smooth dynamics of optimal paths in the stochastic morris-lecar model: Type i and ii excitability. Nonlinear Dyn. 96(4), 2293–2305 (2019)
Dahiya, D., Cameron, M.: Ordered line integral methods for computing the quasi-potential. J. Sci. Comput. 75(3), 1351–1384 (2018)
Dai, M., Gao, T., Lu, Y., Zheng, Y., Duan, J.: Detecting the maximum likelihood transition path from data of stochastic dynamical systems. Chaos 30, 113124 (2020)
Dykman, M.I., Millonas, M.M., Smelyanskiy, V.N.: Observable and hidden singular features of large fluctuations in nonequilibrium systems. Phys. Lett. A 195(1), 53–58 (1994)
Dykman, M.I., Mori, E., Ross, J., Hunt, P.: Large fluctuations and optimal paths in chemical kinetics. J. Chem. Phys. 100(8), 5735–5750 (1994)
Ge, H., Qian, H.: Analytical mechanics in stochastic dynamics: most probable path, large-deviation rate function and hamilton-jacobi equation. Int. J. Modern Phys. B 26(24), 1230012 (2012)
Grafke, T., Vanden-Eijnden, E.: Numerical computation of rare events via large deviation theory. Chaos 29(6), 063118 (2019)
Heymann, M., Vanden-Eijnden, E.: The geometric minimum action method: a least action principle on the space of curves. Commun. Pure Appl. Math. 61(8), 1052–1117 (2008)
Iswarya, M., Raja, R., Rajchakit, G., Cao, J., Alzabut, J., Huang, C.: A perspective on graph theory-based stability analysis of impulsive stochastic recurrent neural networks with time-varying delays. Adv. Diff. Equ. 2019(1), 1–21 (2019)
Klus, S., Nüske, F., Peitz, S., Niemann, J.H., Clementi, C., Schütte, C.: Data-driven approximation of the koopman generator: Model reduction, system identification, and control. Physica D: Nonlinear Phenomena 406, 132416 (2020)
Li, Y., Duan, J.: A data-driven approach for discovering stochastic dynamical systems with non-gaussian lévy noise. Physica D: Nonlinear Phenomena 417, 132830 (2021)
Li, Y., Duan, J., Liu, X.: Machine learning framework for computing the most probable paths of stochastic dynamical systems. Phys. Rev. E 103(1), 012124 (2021)
Li, Y., Liu, X.: Noise induced escape in one-population and two-population stochastic neural networks with internal states. Chaos 29(2), 023137 (2019)
Lin, B. Li, Q., Ren, W.: A data driven method for computing quasipotentials. arXiv preprint arXiv:2012.09111, 2020
Freidlin, M.I., Wentzell, A.D.: Random perturbations of dynamical systems. Springer, Berlin (2012)
Maier, R.S., Stein, D.L.: Effect of focusing and caustics on exit phenomena in systems lacking detailed balance. Phys. Rev. Lett. 71(12), 1783 (1993)
Maier, R.S., Stein, D.L.: A scaling theory of bifurcations in the symmetric weak-noise escape problem. J. Stat. Phys. 83(3), 291–357 (1996)
Manickam, I., Ramachandran, R., Rajchakit, G., Cao, J., Huang, C.: Novel lagrange sense exponential stability criteria for time-delayed stochastic cohen-grossberg neural networks with markovian jump parameters: a graph-theoretic approach. Nonlinear Anal.: Model. Control 25(5), 726–744 (2020)
Raissi, M., Perdikaris, P., Karniadakis, G.: Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. J. Comput. Phys. 378, 686–707 (2019)
Smelyanskiy, V., Dykman, M., Maier, R.: Topological features of large fluctuations to the interior of a limit cycle. Phys. Rev. E 55(3), 2369 (1997)
Tindjong, R., Kaufman, I., Luchinsky, D.G., McClintock, P.V., Khovanov, I.A., Eisenberg, R.: Non-equilibrium stochastic dynamics of open ion channels. Nonlinear Phenom. Complex Syst. 16(2), 146–161 (2013)
Yang X., Huang C., Yang Z.: Stochastic synchronization of reaction-diffusion neural networks under general impulsive controller with mixed delays. In: Abstract and applied analysis, volume 2012. Hindawi, 2012
Zhang, Y., Duan, J., Jin, Y., Li, Y.: Extracting non-gaussian governing laws from data on mean exit time. Chaos 30(11), 113112 (2020)
Zheng, Y., Yang, F., Duan, J., Sun, X., Fu, L., Kurths, J.: The maximum likelihood climate change for global warming under the influence of greenhouse effect and lévy noise. Chaos 30, 013132 (2020)
Zhou, X., Ren, W., Weinan, E.: Adaptive minimum action method for the study of rare events. J. Chem. Phys. 128(10), 104111 (2008)
Zhou, Y., Wan, X., Huang, C., Yang, X.: Finite-time stochastic synchronization of dynamic networks with nonlinear coupling strength via quantized intermittent control. Appl. Math. Comput. 376, 125157 (2020)
Zhu, W., Wu, Y.: First-passage time of duffing oscillator under combined harmonic and white-noise excitations. Nonlinear Dyn. 32(3), 291–305 (2003)
Acknowledgements
We would like to thank Xiaoli Chen for helpful discussions.
Funding
This work was supported by the NSFC 62073166, 61673215, the 333 Project, and the Key Laboratory of Jiangsu Province.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of Interest
The authors declare that they have no conflict of interest.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Li, Y., Xu, S., Duan, J. et al. A machine learning method for computing quasi-potential of stochastic dynamical systems. Nonlinear Dyn 109, 1877–1886 (2022). https://doi.org/10.1007/s11071-022-07536-x
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11071-022-07536-x