{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,13]],"date-time":"2026-06-13T01:57:23Z","timestamp":1781315843384,"version":"3.54.1"},"reference-count":58,"publisher":"AIP Publishing","issue":"10","funder":[{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["No. 12472033"],"award-info":[{"award-number":["No. 12472033"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["pubs.aip.org"],"crossmark-restriction":true},"short-container-title":[],"published-print":{"date-parts":[[2025,10,1]]},"abstract":"<jats:p>The prediction of high-dimensional chaotic time series is a widely concerned and highly challenging task. We propose a novel method, reservoir computing Kolmogorov\u2013Arnold networks, that combines reservoir computing and Kolmogorov\u2013Arnold networks to address this difficult issue. The proposed method involves replacing the linear output layer of traditional reservoir computing with Kolmogorov\u2013Arnold networks and employs a training scheme that incorporates the teacher-forcing mode and the free-running mode during the training process, respectively. We utilize the multiscale Lorenz-96 system and the Kuramoto\u2013Sivashinsky equation as two examples of high-dimensional chaotic systems to illustrate the feasibility of our method. The results of comparative experiments demonstrate that our method exhibits superior prediction capabilities and maintains robustness in the presence of noise perturbation and hyperparameter variations.<\/jats:p>","DOI":"10.1063\/5.0289577","type":"journal-article","created":{"date-parts":[[2025,10,15]],"date-time":"2025-10-15T12:57:09Z","timestamp":1760533029000},"update-policy":"https:\/\/doi.org\/10.1063\/aip-crossmark-policy-page","source":"Crossref","is-referenced-by-count":1,"title":["Reservoir computing meeting Kolmogorov\u2013Arnold networks: Prediction of high-dimensional chaotic systems"],"prefix":"10.1063","volume":"35","author":[{"ORCID":"https:\/\/orcid.org\/0009-0006-4624-8177","authenticated-orcid":false,"given":"Gen","family":"Li","sequence":"first","affiliation":[{"name":"School of Mathematics and Statistics, Northwestern Polytechnical University 1 , Xi\u2019an 710072,","place":["China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0004-2725-637X","authenticated-orcid":false,"given":"Liang","family":"Huang","sequence":"additional","affiliation":[{"name":"School of Mathematics and Statistics, Northwestern Polytechnical University 1 , Xi\u2019an 710072,","place":["China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8078-1067","authenticated-orcid":false,"given":"Youming","family":"Lei","sequence":"additional","affiliation":[{"name":"School of Mathematics and Statistics, Northwestern Polytechnical University 1 , Xi\u2019an 710072,","place":["China"]},{"name":"Ministry of Industry and Information Technology Key Laboratory of Dynamics and Control of Complex Systems, Northwestern Polytechnical University 2 , Xi\u2019an 710072,","place":["China"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"317","published-online":{"date-parts":[[2025,10,15]]},"reference":[{"key":"2025101508483543500_c1","doi-asserted-by":"publisher","first-page":"749","DOI":"10.1016\/j.chaos.2003.12.022","article-title":"A symmetric image encryption scheme based on 3D chaotic cat maps","volume":"21","year":"2004","journal-title":"Chaos, Solitons Fractals"},{"key":"2025101508483543500_c2","doi-asserted-by":"publisher","first-page":"141103","DOI":"10.1103\/PhysRevLett.120.141103","article-title":"Matching matched filtering with deep networks for gravitational-wave astronomy","volume":"120","year":"2018","journal-title":"Phys. Rev. Lett."},{"key":"2025101508483543500_c3","doi-asserted-by":"publisher","first-page":"132911","DOI":"10.1016\/j.physd.2021.132911","article-title":"Supervised learning from noisy observations: Combining machine-learning techniques with data assimilation","volume":"423","year":"2021","journal-title":"Physica D"},{"key":"2025101508483543500_c4","doi-asserted-by":"publisher","first-page":"311","DOI":"10.1016\/S0960-0779(02)00408-3","article-title":"Complex dynamics of Holling type II Lotka\u2013Volterra predator\u2013prey system with impulsive perturbations on the predator","volume":"16","year":"2003","journal-title":"Chaos, Solitons Fractals"},{"key":"2025101508483543500_c5"},{"key":"2025101508483543500_c6","unstructured":"Y.\n            Lei\n          , J.Hu, and J.Ding, \u201cA hybrid model based on deep LSTM for predicting high-dimensional chaotic systems,\u201d arXiv:2002.00799 (2020)."},{"key":"2025101508483543500_c7","doi-asserted-by":"publisher","first-page":"110045","DOI":"10.1016\/j.chaos.2020.110045","article-title":"Robustness of LSTM neural networks for multi-step forecasting of chaotic time series","volume":"139","year":"2020","journal-title":"Chaos, Solitons Fractals"},{"key":"2025101508483543500_c8","doi-asserted-by":"publisher","first-page":"111570","DOI":"10.1016\/j.chaos.2021.111570","article-title":"Forecasting of noisy chaotic systems with deep neural networks","volume":"153","year":"2021","journal-title":"Chaos, Solitons Fractals"},{"key":"2025101508483543500_c9","doi-asserted-by":"publisher","first-page":"041102","DOI":"10.1063\/1.4979665","article-title":"Reservoir observers: Model-free inference of unmeasured variables in chaotic systems","volume":"27","year":"2017","journal-title":"Chaos"},{"key":"2025101508483543500_c10","doi-asserted-by":"publisher","first-page":"121102","DOI":"10.1063\/1.5010300","article-title":"Using machine learning to replicate chaotic attractors and calculate Lyapunov exponents from data","volume":"27","year":"2017","journal-title":"Chaos"},{"key":"2025101508483543500_c11","doi-asserted-by":"publisher","first-page":"061104","DOI":"10.1063\/1.5039508","article-title":"Attractor reconstruction by machine learning","volume":"28","year":"2018","journal-title":"Chaos"},{"key":"2025101508483543500_c12","doi-asserted-by":"publisher","first-page":"2531","DOI":"10.1162\/089976602760407955","article-title":"Real-time computing without stable states: A new framework for neural computation based on perturbations","volume":"14","year":"2002","journal-title":"Neural Comput."},{"key":"2025101508483543500_c13","doi-asserted-by":"publisher","first-page":"78","DOI":"10.1126\/science.1091277","article-title":"Harnessing nonlinearity: Predicting chaotic systems and saving energy in wireless communication","volume":"304","year":"2004","journal-title":"Science"},{"key":"2025101508483543500_c14","doi-asserted-by":"publisher","first-page":"127","DOI":"10.1016\/j.cosrev.2009.03.005","article-title":"Reservoir computing approaches to recurrent neural network training","volume":"3","year":"2009","journal-title":"Comput. Sci. Rev."},{"key":"2025101508483543500_c15","doi-asserted-by":"publisher","first-page":"123108","DOI":"10.1063\/1.5120710","article-title":"Forecasting chaotic systems with very low connectivity reservoir computers","volume":"29","year":"2019","journal-title":"Chaos"},{"key":"2025101508483543500_c16","doi-asserted-by":"publisher","first-page":"373","DOI":"10.5194\/npg-27-373-2020","article-title":"Data-driven predictions of a multiscale Lorenz 96 chaotic system using machine-learning methods: Reservoir computing, artificial neural network, and long short-term memory network","volume":"27","year":"2020","journal-title":"Nonlinear Process. Geophys."},{"key":"2025101508483543500_c17","doi-asserted-by":"publisher","first-page":"5564","DOI":"10.1038\/s41467-021-25801-2","article-title":"Next generation reservoir computing","volume":"12","year":"2021","journal-title":"Nat. Commun."},{"key":"2025101508483543500_c18","doi-asserted-by":"publisher","first-page":"191","DOI":"10.1016\/j.neunet.2020.02.016","article-title":"Backpropagation algorithms and reservoir computing in recurrent neural networks for the forecasting of complex spatiotemporal dynamics","volume":"126","year":"2020","journal-title":"Neural Netw."},{"key":"2025101508483543500_c19","doi-asserted-by":"publisher","first-page":"164101","DOI":"10.1103\/PhysRevLett.128.164101","article-title":"Parallel machine learning for forecasting the dynamics of complex networks","volume":"128","year":"2022","journal-title":"Phys. Rev. Lett."},{"key":"2025101508483543500_c20","doi-asserted-by":"publisher","first-page":"530","DOI":"10.1016\/j.neunet.2022.06.025","article-title":"A systematic exploration of reservoir computing for forecasting complex spatiotemporal dynamics","volume":"153","year":"2022","journal-title":"Neural Netw."},{"key":"2025101508483543500_c21","doi-asserted-by":"publisher","first-page":"041101","DOI":"10.1063\/1.5028373","article-title":"Hybrid forecasting of chaotic processes: Using machine learning in conjunction with a knowledge-based model","volume":"28","year":"2018","journal-title":"Chaos"},{"key":"2025101508483543500_c22","doi-asserted-by":"publisher","first-page":"024102","DOI":"10.1103\/PhysRevLett.120.024102","article-title":"Model-free prediction of large spatiotemporally chaotic systems from data: A reservoir computing approach","volume":"120","year":"2018","journal-title":"Phys. Rev. Lett."},{"key":"2025101508483543500_c23","doi-asserted-by":"publisher","first-page":"053111","DOI":"10.1063\/5.0005541","article-title":"Combining machine learning with knowledge-based modeling for scalable forecasting and subgrid-scale closure of large, complex, spatiotemporal systems","volume":"30","year":"2020","journal-title":"Chaos"},{"key":"2025101508483543500_c24","doi-asserted-by":"publisher","first-page":"113325","DOI":"10.1016\/j.chaos.2023.113325","article-title":"Chimera states in coupled pendulum with higher-order interaction","volume":"170","year":"2023","journal-title":"Chaos, Solitons Fractals"},{"key":"2025101508483543500_c25","doi-asserted-by":"publisher","first-page":"064304","DOI":"10.1103\/PhysRevE.108.064304","article-title":"Reservoir computing with higher-order interactive coupled pendulums","volume":"108","year":"2023","journal-title":"Phys. Rev. E"},{"key":"2025101508483543500_c26","doi-asserted-by":"publisher","first-page":"073109","DOI":"10.1063\/5.0152311","article-title":"Data-informed reservoir computing for efficient time-series prediction","volume":"33","year":"2023","journal-title":"Chaos"},{"key":"2025101508483543500_c27","doi-asserted-by":"publisher","first-page":"033111","DOI":"10.1063\/5.0138661","article-title":"Reservoir computing as digital twins for nonlinear dynamical systems","volume":"33","year":"2023","journal-title":"Chaos"},{"key":"2025101508483543500_c28","doi-asserted-by":"publisher","first-page":"024227","DOI":"10.1103\/PhysRevE.109.024227","article-title":"Constructing polynomial libraries for reservoir computing in nonlinear dynamical system forecasting","volume":"109","year":"2024","journal-title":"Phys. Rev. E"},{"key":"2025101508483543500_c29","doi-asserted-by":"publisher","first-page":"063114","DOI":"10.1063\/5.0206232","article-title":"Hybridizing traditional and next-generation reservoir computing to accurately and efficiently forecast dynamical systems","volume":"34","year":"2024","journal-title":"Chaos"},{"key":"2025101508483543500_c30","doi-asserted-by":"publisher","first-page":"2506","DOI":"10.1038\/s41467-024-46852-1","article-title":"Higher-order Granger reservoir computing: Simultaneously achieving scalable complex structures inference and accurate dynamics prediction","volume":"15","year":"2024","journal-title":"Nat. Commun."},{"key":"2025101508483543500_c31","doi-asserted-by":"publisher","first-page":"5961","DOI":"10.1038\/s41467-025-61195-1","article-title":"Learning dynamical systems with hit-and-run random feature maps","volume":"16","year":"2025","journal-title":"Nat. Commun."},{"key":"2025101508483543500_c32","doi-asserted-by":"publisher","first-page":"115137","DOI":"10.1016\/j.chaos.2024.115137","article-title":"Prediction of spatiotemporal dynamic systems by data-driven reconstruction","volume":"185","year":"2024","journal-title":"Chaos, Solitons Fractals"},{"key":"2025101508483543500_c33","unstructured":"J.\n            Xu\n          , Z.Chen, J.Li, S.Yang, W.Wang, X.Hu, and E. C.-H.Ngai, \u201cFourierKAN-GCF: Fourier Kolmogorov-Arnold network\u2014An effective and efficient feature transformation for graph collaborative filtering,\u201d arXiv:2406.01034 (2024)."},{"key":"2025101508483543500_c34","unstructured":"Y.\n            Xu\n          , G. A.Gottwald, and Z.Kuncic, \u201cLearning chaotic dynamics with neuromorphic network dynamics,\u201d arXiv:2506.10773 (2025)."},{"key":"2025101508483543500_c35","doi-asserted-by":"publisher","first-page":"106575","DOI":"10.1016\/j.neunet.2024.106575","article-title":"Photonic deep residual time-delay reservoir computing","volume":"179","year":"2024","journal-title":"Neural Netw."},{"key":"2025101508483543500_c36","unstructured":"D. A.\n            Norton\n          , E.Ott, A.Pomerance, B.Hunt, and M.Girvan, \u201cTailored forecasting from short time series via meta-learning,\u201d arXiv:2501.16325 (2025)."},{"key":"2025101508483543500_c37","doi-asserted-by":"crossref","unstructured":"P.\n            Mandal\n          , G. A.Gottwald, and N.Cranch, \u201cOn the choice of the non-trainable internal weights in random feature maps,\u201d arXiv:2408.03626 (2024).","DOI":"10.3934\/fods.2025006"},{"key":"2025101508483543500_c38","unstructured":"Z.\n            Liu\n          , Y.Wang, S.Vaidya, F.Ruehle, J.Halverson, M.Solja\u010di\u0107, T. Y.Hou, and M.Tegmark, \u201cKAN: Kolmogorov-Arnold networks,\u201d arXiv:2404.19756 (2024)."},{"key":"2025101508483543500_c39","unstructured":"Z.\n            Liu\n          , P.Ma, Y.Wang, W.Matusik, and M.Tegmark, \u201cKAN 2.0: Kolmogorov-Arnold networks meet science,\u201d arXiv:2408.10205 (2024)."},{"key":"2025101508483543500_c40","doi-asserted-by":"crossref","unstructured":"R.\n            Genet\n           and H.Inzirillo, \u201cTKAN: Temporal Kolmogorov-Arnold networks,\u201d arXiv:2405.07344 (2024).","DOI":"10.2139\/ssrn.4825654"},{"key":"2025101508483543500_c41","doi-asserted-by":"crossref","unstructured":"R.\n            Genet\n           and H.Inzirillo, \u201cA temporal Kolmogorov-Arnold transformer for time series forecasting,\u201d arXiv:2406.02486 (2024).","DOI":"10.2139\/ssrn.5176911"},{"key":"2025101508483543500_c42","doi-asserted-by":"crossref","unstructured":"H.\n            Inzirillo\n           and R.Genet, \u201cSigKAN: Signature-weighted Kolmogorov-Arnold networks for time series,\u201d arXiv:2406.17890 (2024).","DOI":"10.2139\/ssrn.5183808"},{"key":"2025101508483543500_c43","unstructured":"H.\n            Inzirillo\n           and R.Genet, \u201cA gated residual Kolmogorov-Arnold networks for mixtures of experts,\u201d arXiv:2409.15161 (2024)."},{"key":"2025101508483543500_c44","unstructured":"X.\n            Han\n          , X.Zhang, Y.Wu, Z.Zhang, and Z.Wu, \u201cKAN4TSF: Are KAN and KAN-based models effective for time series forecasting?\u201d arXiv:2408.11306 (2024)."},{"key":"2025101508483543500_c45","unstructured":"H.-T.\n            Ta\n          , D.-Q.Thai, A. B. S.Rahman, G.Sidorov, and A.Gelbukh, \u201cFC-KAN: Function combinations in Kolmogorov-Arnold networks,\u201d arXiv:2409.01763 (2024)."},{"key":"2025101508483543500_c46","doi-asserted-by":"crossref","unstructured":"Z.\n            Bozorgasl\n           and H.Chen, \u201cWav-KAN: Wavelet Kolmogorov-Arnold networks,\u201d arXiv:2405.12832 (2024).","DOI":"10.2139\/ssrn.4835325"},{"key":"2025101508483543500_c47","doi-asserted-by":"publisher","first-page":"023037","DOI":"10.1103\/PhysRevResearch.7.023037","article-title":"Data-driven model discovery with Kolmogorov-Arnold networks","volume":"7","year":"2025","journal-title":"Phys. Rev. Res."},{"key":"2025101508483543500_c48","first-page":"953","article-title":"On the representation of continuous functions of many variables by superposition of continuous functions of one variable and addition","volume":"114","year":"1957","journal-title":"Dokl. Akad. Nauk"},{"key":"2025101508483543500_c49","volume-title":"Proceedings of the 28th International Conference on Neural Information Processing Systems, NIPS\u201915"},{"key":"2025101508483543500_c50","unstructured":"M.\n            Ranzato\n          , S.Chopra, M.Auli, and W.Zaremba, \u201cSequence level training with recurrent neural networks,\u201d arXiv:1511.06732 (2015)."},{"key":"2025101508483543500_c51"},{"key":"2025101508483543500_c52","doi-asserted-by":"publisher","first-page":"100","DOI":"10.1016\/j.physd.2005.09.011","article-title":"Testing for chaos in deterministic systems with noise","volume":"212","year":"2005","journal-title":"Physica D"},{"key":"2025101508483543500_c53","doi-asserted-by":"publisher","first-page":"34","DOI":"10.1016\/j.physd.2013.02.009","article-title":"A statistically accurate modified quasilinear Gaussian closure for uncertainty quantification in turbulent dynamical systems","volume":"252","year":"2013","journal-title":"Physica D"},{"key":"2025101508483543500_c54","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1016\/j.physd.2015.12.008","article-title":"Filter accuracy for the Lorenz 96 model: Fixed versus adaptive observation operators","volume":"325","year":"2016","journal-title":"Physica D"},{"key":"2025101508483543500_c55","doi-asserted-by":"publisher","first-page":"1452","DOI":"10.1007\/s12559-020-09732-6","article-title":"Bayesian optimisation of large-scale photonic reservoir computers","volume":"15","year":"2023","journal-title":"Cogn. Comput."},{"key":"2025101508483543500_c56","doi-asserted-by":"publisher","first-page":"1207","DOI":"10.1016\/0094-5765(77)90097-2","article-title":"Nonlinear analysis of hydrodynamic instability in laminar flames\u2014II. Numerical experiments","volume":"4","year":"1977","journal-title":"Acta Astronaut."},{"key":"2025101508483543500_c57","doi-asserted-by":"publisher","first-page":"346","DOI":"10.1143\/PTPS.64.346","article-title":"Diffusion-induced chaos in reaction systems","volume":"64","year":"1978","journal-title":"Prog. Theor. Phys. Suppl."},{"key":"2025101508483543500_c58"}],"container-title":["Chaos: An Interdisciplinary Journal of Nonlinear Science"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/pubs.aip.org\/aip\/cha\/article-pdf\/doi\/10.1063\/5.0289577\/20754562\/103120_1_5.0289577.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/pubs.aip.org\/aip\/cha\/article-pdf\/doi\/10.1063\/5.0289577\/20754562\/103120_1_5.0289577.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,10,15]],"date-time":"2025-10-15T12:57:13Z","timestamp":1760533033000},"score":1,"resource":{"primary":{"URL":"https:\/\/pubs.aip.org\/cha\/article\/35\/10\/103120\/3367908\/Reservoir-computing-meeting-Kolmogorov-Arnold"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,10,1]]},"references-count":58,"journal-issue":{"issue":"10","published-print":{"date-parts":[[2025,10,1]]}},"URL":"https:\/\/doi.org\/10.1063\/5.0289577","relation":{},"ISSN":["1054-1500","1089-7682"],"issn-type":[{"value":"1054-1500","type":"print"},{"value":"1089-7682","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2025,10]]},"published":{"date-parts":[[2025,10,1]]},"article-number":"103120"}}