Learning dynamics by reservoir computing (In Memory of Prof. Pavol Brunovský). (English) Zbl 1539.37082
Summary: We study reservoir computing, a machine learning method, from the viewpoint of learning dynamics. We present numerical results of learning the dynamics of the logistic map, one of the typical examples of chaotic dynamical systems, using a 30-node reservoir and a three-node reservoir. When the learning is successful, an attractor that is smoothly conjugate to the logistic map to be learned is observed in the phase space of the reservoir. Inspired by this numerical result, we introduce a degenerate reservoir system and use it to mathematically confirm this observation. We also show that reservoir computing can learn information about dynamics not included in the training data, which we believe is a remarkable feature of reservoir computing compared to other machine learning methods. We discuss this feature in connection with the above observation that there is a smooth conjugacy between the attractor in the reservoir and the dynamics to be learned.
MSC:
| 37M05 | Simulation of dynamical systems |
| 37M22 | Computational methods for attractors of dynamical systems |
| 37M20 | Computational methods for bifurcation problems in dynamical systems |
| 37E05 | Dynamical systems involving maps of the interval |
References:
| [1] | Jaeger, H.: The echo state approach to analysing and trining recurrent neural networks, Technical Report, GMD Report 148, (2001), GMD—German National Research Institute for Computer Science. http://minds.jacobs-university.de/pubs; Erratum note (2010) |
| [2] | Jaeger, H.; Haas, H., Harnessing nonlinearity: predicting chaotic systems and saving energy in wireless communication, Science, 304, 78, 2004 · doi:10.1126/science.1091277 |
| [3] | Kobayashi, M.U., Nakai, K., Saiki, Y., Tsutsumi, N.: Dynamical system analysis of a data-driven model constructed by reservoir computing. Phys. Rev. E 104, 044215 (2021). doi:10.1103/PhysRevE.104.044215 |
| [4] | Maass, W.; Natschläger, T.; Markram, H., Real-time computing without stable staes: a new framework for neural computation based on perturbations, Neural Comput., 14, 2531-2560, 2002 · Zbl 1057.68618 · doi:10.1162/089976602760407955 |
| [5] | Pathak, J.; Hunt, BR; Girvan, M.; Lu, Z.; Ott, E., Model-free prediction of large spatiotemporally chaotic systems from data: a reservoir computing approach, Phys. Rev. Lett., 120, 2018 · doi:10.1103/PhysRevLett.120.024102 |
| [6] | Pathak, J.; Lu, Z.; Hunt, BR; Girvan, M.; Ott, E., Using machine learning to replicate chaotic attractors and calculate Lyapunov exponents from data, Chaos, 27, 2017 · Zbl 1390.37138 · doi:10.1063/1.5010300 |
| [7] | Robinson, C.: Dynamical Systems: Stability, Symbolic Dynamics, and Chaos, 2nd edn. CRC Press (1998) |
| [8] | Lu, Z.; Hunt, BR; Ott, E., Attractor reconstruction by machine learning, Chaos, 28, 2018 · doi:10.1063/1.5039508 |
| [9] | Lu, Z.; Pathak, J.; Hunt, BR; Girvan, M.; Brockett, R.; Ott, E., Reservoir observers: model-free inference of unmeasured variables in chaotic systems, Chaos, 27, 2017 · doi:10.1063/1.4979665 |
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.
