arxiv: 2605.13878 · v1 · submitted 2026-05-10 · 🌊 nlin.CD · math.DS· physics.data-an

Recognition: 2 theorem links

· Lean Theorem

Revealing dynamics of non-autonomous complex systems from data

Chengzuo Zhuge , Zheng Jiang , Zhefan Xu , Wei Chen

Authors on Pith no claims yet

Pith reviewed 2026-05-15 05:45 UTC · model grok-4.3

classification 🌊 nlin.CD math.DSphysics.data-an

keywords non-autonomous dynamicsequation discoverydata-driven modelingbasis functionsdynamical systemscomplex systemsprediction from data

0 comments

The pith

A data-driven method discovers governing equations for non-autonomous systems by selecting optimal basis functions from a library.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

This paper presents a computational approach to infer the equations governing non-autonomous dynamical systems directly from data. It identifies an optimal sparse combination of basis functions that reconstructs the system's behavior under simplified prior assumptions about the model structure. The method is tested on synthetic cases including cusp bifurcations and coupled Kuramoto oscillators, then applied to real observations from leaf cellular energy, unmanned aerial vehicle paths, chick-heart cell aggregates, and marine fish communities. If the approach holds, it allows accurate forward prediction of system evolution and extraction of the underlying laws even when external time-varying influences are unknown. This addresses a gap in handling systems where standard autonomous equation discovery fails due to inherent non-autonomicity.

Core claim

Non-autonomous dynamical equations can be inferred from data by identifying an optimal set of basis functions within the model space, enabling reconstruction of complex system behavior under simplified prior specifications. The paper demonstrates this on canonical synthetic systems such as cusp bifurcation and coupled Kuramoto oscillators, and extends it to empirical cases including leaf cellular energy, unmanned aerial vehicle navigation, chick-heart aggregates, and marine fish communities using simple basis function libraries, yielding accurate predictions of evolution and uncovered governing laws.

What carries the argument

Identifying an optimal sparse set of basis functions from a user-supplied library to represent the non-autonomous dynamics.

If this is right

The inferred equations accurately predict future evolution in the tested empirical systems.
Governing laws can be uncovered for diverse real-world systems from the discovered equations.
The approach succeeds with simple basis function libraries across biological, engineering, and ecological examples.
It extends equation discovery to non-autonomous cases where external influences vary with time.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

If the basis library is expanded, the same framework could handle systems with higher-order or more intricate time dependencies not covered in the tested cases.
The method's outputs could feed into downstream tasks like stability analysis or intervention design for the studied systems.
Success on chick-heart aggregates suggests possible extensions to other excitable biological media where external pacing is present.

Load-bearing premise

The true non-autonomous dynamics can be represented accurately as a sparse linear combination of functions from a supplied basis library, with the optimal subset identifiable under simplified priors.

What would settle it

Applying the method to any of the real systems, such as the marine fish community, and finding that the inferred equations produce large prediction errors on held-out time series would falsify the claim.

read the original abstract

Discovering governing equations from data is crucial for understanding complex systems in many diverse fields from science to engineering. Yet, there still is a lack of versatile computational toolbox to deal with this long standing challenge due to the inherent non-autonomicity and unknowability of the underlying dynamics. Here, we introduce a data-driven approach for inferring non-autonomous dynamical equations by identifying an optimal set of basis functions within the model space, enabling the reconstruction of complex systems behavior under simplified prior specifications. Our method demonstrates effectiveness in equation discovery on canonical synthetic systems such as cusp bifurcation and coupled Kuramoto oscillators. Furthermore, we extend the application of this approach to leaf cellular energy, unmanned aerial vehicle navigation, chick-heart aggregates, and marine fish community under simple basis function libraries. Leveraging the inferred equations, we accurately predict the evolution of these empirical systems and further uncover their governing laws. Our approach offers a novel paradigm to reveal the underlying dynamics of a wide range of real-world systems.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

This paper gives a practical basis-selection method for equation discovery in non-autonomous systems and shows it works on both standard tests and four real datasets, though success still depends on the user supplying a good library.

read the letter

The main contribution is a procedure that picks an optimal subset of basis functions to recover time-varying governing equations from data. It takes the sparse-regression framework and adds an explicit step for handling non-autonomy under simplified priors. The synthetic tests recover the cusp bifurcation and coupled Kuramoto dynamics, and the real examples cover leaf cellular energy, UAV navigation, chick-heart aggregates, and marine fish communities. In each case the inferred equations are used to predict evolution and extract apparent laws. The demonstrations are concrete and the domains are diverse, which makes the work immediately usable for applied modeling. The algorithmic description appears clear enough that someone could implement the selection step without major gaps. The stress-test note confirms that the library assumption is stated upfront and checked on systems where it holds, with no hidden inconsistencies in the argument. The soft spot is the continued reliance on a user-supplied library: if the true vector field lies outside that span, the method will either fail to fit or produce misleading terms. The abstract gives no error bars or cross-validation numbers, but the full text supplies benchmarks on known systems, so the central claim is at least testable. This paper is for researchers already working with data-driven equation discovery who need to move beyond autonomous assumptions. A reader looking for a ready-to-apply tool on time-varying problems will find the case studies useful. It deserves a serious referee because the examples are broad enough to expose practical strengths and limits, even if the theoretical advance is incremental.

Referee Report

2 major / 1 minor

Summary. The paper introduces a data-driven method for inferring non-autonomous dynamical equations by selecting an optimal subset of basis functions from a user-supplied library under simplified priors. It validates the approach on synthetic benchmarks (cusp bifurcation and coupled Kuramoto oscillators) and applies it to four real-world datasets (leaf cellular energy, UAV navigation, chick-heart aggregates, and marine fish communities), claiming that the resulting sparse models accurately predict system evolution and reveal governing laws.

Significance. If the central claim holds under rigorous validation, the method would supply a practical basis-selection procedure for non-autonomous equation discovery, extending sparse regression techniques to time-dependent systems. The explicit use of enumerated libraries on both canonical synthetic cases and diverse empirical examples is a strength, as is the emphasis on predictive accuracy rather than purely descriptive fitting. These elements could make the work a useful addition to the nlin.CD literature on data-driven modeling.

major comments (2)

[Abstract] Abstract: the effectiveness claims for both synthetic and real examples are stated without any quantitative metrics (prediction error, R², cross-validation scores, or time-dependent error bars), so the central assertion that the method 'accurately predict[s] the evolution' and 'uncover[s] governing laws' cannot be assessed from the given text.
[Methods] The weakest assumption—that the true vector field lies in the span of the supplied basis library—is load-bearing for all results; the manuscript must demonstrate that library choice is independent of the fitting data (e.g., via hold-out or a priori specification) to rule out circularity in the selection step.

minor comments (1)

[Applications] Results sections on the four empirical examples should list the exact basis functions employed for each dataset to permit reproducibility.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback. We have revised the manuscript to incorporate quantitative metrics throughout and to explicitly demonstrate the a priori, data-independent selection of basis libraries via hold-out validation.

read point-by-point responses

Referee: [Abstract] Abstract: the effectiveness claims for both synthetic and real examples are stated without any quantitative metrics (prediction error, R², cross-validation scores, or time-dependent error bars), so the central assertion that the method 'accurately predict[s] the evolution' and 'uncover[s] governing laws' cannot be assessed from the given text.

Authors: We agree that the abstract would be strengthened by quantitative support. In the revised manuscript we have updated the abstract to report average prediction errors and R² scores for the synthetic benchmarks and the four empirical examples. We have also expanded the results and supplementary sections with full cross-validation scores, time-dependent error bars, and hold-out prediction metrics to substantiate the claims of accurate evolution prediction and law discovery. revision: yes
Referee: [Methods] The weakest assumption—that the true vector field lies in the span of the supplied basis library—is load-bearing for all results; the manuscript must demonstrate that library choice is independent of the fitting data (e.g., via hold-out or a priori specification) to rule out circularity in the selection step.

Authors: We thank the referee for identifying this critical point. The libraries were selected a priori from simplified, domain-informed priors (e.g., polynomial and trigonometric terms for oscillators, energy-related terms for biological systems) without reference to the fitting data. To rigorously exclude circularity we have added a dedicated hold-out validation subsection: each dataset is partitioned into training and test portions, the library is fixed using only the training portion, and predictive performance is reported on the unseen test data, confirming that library choice remains independent of the data used for final fitting. revision: yes

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The paper introduces a sparse basis-function selection procedure for inferring non-autonomous dynamics from data, with the core assumption (true vector field lies in the span of a user-supplied library) stated explicitly and tested on synthetic systems whose ground-truth equations are known independently (cusp bifurcation, coupled Kuramoto). Recovery on these benchmarks is verifiable against external truth rather than tautological. Empirical applications use the same procedure to produce predictive models whose accuracy is assessed on held-out evolution, without any self-definitional loop, fitted parameter renamed as prediction, or load-bearing self-citation chain. No uniqueness theorems, ansatzes imported via citation, or renaming of known results appear. The derivation chain is therefore self-contained and relies on standard sparse regression applied to time-series data.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The central claim rests on the assumption that non-autonomous dynamics lie within the span of a chosen basis-function library and that an optimal sparse subset can be recovered from data alone.

free parameters (1)

basis function library
The specific library of candidate functions is supplied by the user and directly determines which equations can be discovered.

axioms (1)

domain assumption System dynamics can be expressed as a sparse linear combination of basis functions from a predefined library
Invoked when the method identifies an optimal set of basis functions to reconstruct behavior.

pith-pipeline@v0.9.0 · 5472 in / 1405 out tokens · 52654 ms · 2026-05-15T05:45:36.652640+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

Theorem. For all ν1, ∆ν ∈ R, span{ψi(x,Φ)} = span{ψj(x,ν)}, Col(ΘΦ) = Col(Θν).
IndisputableMonolith/Foundation/AlphaCoordinateFixation.lean alpha_pin_under_high_calibration unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

We employ the grid search in conjunction with the numerical error adjusted Akaike’s information criterion (εAIC) to obtain the optimal pair (ν1, ∆ν)

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

66 extracted references · 66 canonical work pages

[1]

De Col, V.et al.Atp sensing in living plant cells reveals tissue gradients and stress dynamics of energy physiology.Elife6, e26770 (2017). 21

work page 2017
[2]

Wagner, S.et al.Multiparametric real-time sensing of cytosolic physiology links hypoxia responses to mitochondrial electron transport.New Phytologist224, 1668– 1684 (2019)

work page 2019
[3]

Influence of water flow, water temperature and light on fish migration in rivers.Nordic Journal of Freshwater Research66, 20–35 (1991)

Jonsson, N. Influence of water flow, water temperature and light on fish migration in rivers.Nordic Journal of Freshwater Research66, 20–35 (1991)

work page 1991
[4]

P., Francis, C

Marra, P. P., Francis, C. M., Mulvihill, R. S. & Moore, F. R. The influence of climate on the timing and rate of spring bird migration.Oecologia142, 307–315 (2005)

work page 2005
[5]

In IEEE Intelligent Vehicles Symposium, 163–168 (IEEE, 2011)

Levinson, J.et al.Towards fully autonomous driving: Systems and algorithms. In IEEE Intelligent Vehicles Symposium, 163–168 (IEEE, 2011)

work page 2011
[6]

& Takeda, K

Yurtsever, E., Lambert, J., Carballo, A. & Takeda, K. A survey of autonomous driving: Common practices and emerging technologies.IEEE Access8, 58443– 58469 (2020)

work page 2020
[7]

L., Proctor, J

Brunton, S. L., Proctor, J. L. & Kutz, J. N. Discovering governing equations from data by sparse identification of nonlinear dynamical systems.Proceedings of the National Academy of Sciences113, 3932–3937 (2016)

work page 2016
[8]

M., Brunton, S

Mangan, N. M., Brunton, S. L., Proctor, J. L. & Kutz, J. N. Inferring biological networks by sparse identification of nonlinear dynamics.IEEE Transactions on Molecular, Biological, and Multi-Scale Communications2, 52–63 (2017)

work page 2017
[9]

Li, S.et al.Discovering time-varying aerodynamics of a prototype bridge by sparse identification of nonlinear dynamical systems.Physical Review E100, 022220 (2019)

work page 2019
[10]

N., Brunton, B

Fasel, U., Kutz, J. N., Brunton, B. W. & Brunton, S. L. Ensemble-sindy: Robust sparse model discovery in the low-data, high-noise limit, with active learning and control.Proceedings of the Royal Society A478, 20210904 (2022)

work page 2022
[11]

& Eroglu, D

Topal, I. & Eroglu, D. Reconstructing network dynamics of coupled discrete chaotic units from data.Physical Review Letters130, 117401 (2023)

work page 2023
[12]

& Nair, P

Course, K. & Nair, P. B. State estimation of a physical system with unknown governing equations.Nature622, 261–267 (2023)

work page 2023
[13]

& Yan, G

Gao, T.-T. & Yan, G. Autonomous inference of complex network dynamics from incomplete and noisy data.Nature Computational Science2, 160–168 (2022)

work page 2022
[14]

& Yan, G

Gao, T.-T., Barzel, B. & Yan, G. Learning interpretable dynamics of stochas- tic complex systems from experimental data.Nature Communications15, 6029 (2024)

work page 2024
[15]

Li, Z.et al.Bi-level identification of governing equations for nonlinear physical systems.Nature Computational Science5, 456–466 (2025). 22

work page 2025
[16]

& Juniper, M

Fung, L., Fasel, U. & Juniper, M. Rapid bayesian identification of sparse nonlinear dynamics from scarce and noisy data.Proceedings of the Royal Society A481, 20240200 (2025)

work page 2025
[17]

Ushio, M.et al.Fluctuating interaction network and time-varying stability of a natural fish community.Nature554, 360–363 (2018)

work page 2018
[18]

J., van Nes, E

Lever, J. J., van Nes, E. H., Scheffer, M. & Bascompte, J. The sudden collapse of pollinator communities.Ecology Letters17, 350–359 (2014)

work page 2014
[19]

Jiang, J.et al.Predicting tipping points in mutualistic networks through dimen- sion reduction.Proceedings of the National Academy of Sciences115, E639–E647 (2018)

work page 2018
[20]

P., Saavedra, S

Rohr, R. P., Saavedra, S. & Bascompte, J. On the structural stability of mutualistic systems.Science345, 1253497 (2014)

work page 2014
[21]

& Gao, J

Zhang, H., Wang, Q., Zhang, W., Havlin, S. & Gao, J. Estimating comparable distances to tipping points across mutualistic systems by scaled recovery rates. Nature Ecology & Evolution6, 1524–1536 (2022)

work page 2022
[22]

M., Sornette, D

Chen, W., Schr¨ oder, M., D’Souza, R. M., Sornette, D. & Nagler, J. Microtransition cascades to percolation.Physical Review Letters112, 155701 (2014)

work page 2014
[23]

Clemson, P. T. & Stefanovska, A. Discerning non-autonomous dynamics.Physics Reports542, 297–368 (2014)

work page 2014
[24]

& de Valpine, P

Knape, J. & de Valpine, P. Effects of weather and climate on the dynamics of ani- mal population time series.Proceedings of the Royal Society B: Biological Sciences 278, 985–992 (2011)

work page 2011
[25]

Gamelon, M.et al.Interactions between demography and environmental effects are important determinants of population dynamics.Science Advances3, e1602298 (2017)

work page 2017
[26]

Scheffer, M.Critical transitions in nature and society(Princeton University Press, 2020)

work page 2020
[27]

Sornette, D.Critical phenomena in natural sciences: chaos, fractals, selforganiza- tion and disorder: concepts and tools(Springer Science & Business Media, 2006)

work page 2006
[28]

M., Lytle, D

Hern´ andez-Carrasco, D., Tylianakis, J. M., Lytle, D. A. & Tonkin, J. D. Ecological and evolutionary consequences of changing seasonality.Science388, eads4880 (2025)

work page 2025
[29]

Liu, Z.et al.The seasonal temperature conundrum for the holocene.Science Advances11, eadt8950 (2025)

work page 2025
[30]

A.Nonlinear System Identification(John Wiley & Sons, 2013)

Billings, S. A.Nonlinear System Identification(John Wiley & Sons, 2013). 23

work page 2013
[31]

Z., Lu, Z., Nozari, E., Pappas, G

Kim, J. Z., Lu, Z., Nozari, E., Pappas, G. J. & Bassett, D. S. Teaching recurrent neural networks to infer global temporal structure from local examples.Nature Machine Intelligence3, 316–323 (2021)

work page 2021
[32]

J., Bollt, E., Griffith, A

Gauthier, D. J., Bollt, E., Griffith, A. & Barbosa, W. A. Next generation reservoir computing.Nature Communications12, 1–8 (2021)

work page 2021
[33]

A., Kuznetsov, I

Kuznetsov, Y. A., Kuznetsov, I. A. & Kuznetsov, Y.Elements of Applied Bifur- cation Theory(Springer, 1998)

work page 1998
[34]

Self-entrainment of a population of coupled non-linear oscillators

Kuramoto, Y. Self-entrainment of a population of coupled non-linear oscillators. In International Symposium on Mathematical Problems in Theoretical Physics, 420– 422 (1975)

work page 1975
[35]

M.et al.Predicting discrete-time bifurcations with deep learning.Nature Communications14, 6331 (2023)

Bury, T. M.et al.Predicting discrete-time bifurcations with deep learning.Nature Communications14, 6331 (2023)

work page 2023
[36]

& Van Mieghem, P

Prasse, B. & Van Mieghem, P. Predicting network dynamics without requiring the knowledge of the interaction graph.Proceedings of the National Academy of Sciences119, e2205517119 (2022)

work page 2022
[37]

M., Dumais, J

Forterre, Y., Skotheim, J. M., Dumais, J. & Mahadevan, L. How the venus flytrap snaps.Nature433, 421–425 (2005)

work page 2005
[38]

& Ruina, A

Smith, M., Yanega, G. & Ruina, A. Elastic instability model of rapid beak closure in hummingbirds.Journal of Theoretical Biology282, 41–51 (2011)

work page 2011
[39]

Liu, Z.et al.Early predictor for the onset of critical transitions in networked dynamical systems.Physical Review X14, 031009 (2024)

work page 2024
[40]

Champion, K., Lusch, B., Kutz, J. N. & Brunton, S. L. Data-driven discovery of coordinates and governing equations.Proceedings of the National Academy of Sciences116, 22445–22451 (2019)

work page 2019
[41]

Imamura, H.et al.Visualization of atp levels inside single living cells with fluores- cence resonance energy transfer-based genetically encoded indicators.Proceedings of the National Academy of Sciences106, 15651–15656 (2009)

work page 2009
[42]

Doedel, E. J. & Oldeman, B. Auto-07p: Continuation and bifurcation software for ordinary differential equations. Tech. Rep., Concordia University, Montreal (2021)

work page 2021
[43]

InIEEE International Conference on Robotics and Automation (ICRA), 10645–10651 (2023)

Xu, Z.et al.A real-time dynamic obstacle tracking and mapping system for uav navigation and collision avoidance with an rgb-d camera. InIEEE International Conference on Robotics and Automation (ICRA), 10645–10651 (2023)

work page 2023
[44]

& Shimada, K

Xu, Z., Xiu, Y., Zhan, X., Chen, B. & Shimada, K. Vision-aided uav navigation and dynamic obstacle avoidance using gradient-based b-spline trajectory optimization. InIEEE International Conference on Robotics and Automation (ICRA), 1214–1220 (2023). 24

work page 2023
[45]

R., Kristof, A

Clay, J. R., Kristof, A. S., Shenasa, J., Brochu, R. M. & Shrier, A. A review of the effects of three cardioactive agents on the electrical activity from embryonic chick heart cell aggregates: Ttx, ach, and e-4031.Progress in Biophysics and Molecular Biology62, 185–202 (1994)

work page 1994
[46]

Kim, M.-Y.et al.Stochastic and spatial influences on drug-induced bifurcations in cardiac tissue culture.Physical Review Letters103, 058101 (2009)

work page 2009
[47]

Riva, F.et al.Toward a cohesive understanding of ecological complexity.Science Advances9, eabq4207 (2023)

work page 2023
[48]

& Payan, P

Bœuf, G. & Payan, P. How should salinity influence fish growth?Comparative Biochemistry and Physiology Part C: Toxicology & Pharmacology130, 411–423 (2001)

work page 2001
[49]

& Yodzis, P

McCann, K. & Yodzis, P. Nonlinear dynamics and population disappearances. The American Naturalist144, 873–879 (1994)

work page 1994
[50]

Chaotic dynamics of a discrete predator-prey system with prey refuge

Jana, D. Chaotic dynamics of a discrete predator-prey system with prey refuge. Applied Mathematics and Computation224, 848–865 (2013)

work page 2013
[51]

Bifurcations and dynamics of a discrete predator-prey system.Journal of Biological Dynamics8, 161–186 (2014)

Asheghi, R. Bifurcations and dynamics of a discrete predator-prey system.Journal of Biological Dynamics8, 161–186 (2014)

work page 2014
[52]

Levin, S. A. Population dynamic models in heterogeneous environments.Annual Review of Ecology and Systematics7, 287–310 (1976)

work page 1976
[53]

Heterogeneity in dynamic species abundance models: the selective effect of extinction processes.Mathematical Biosciences210, 490–507 (2007)

Engen, S. Heterogeneity in dynamic species abundance models: the selective effect of extinction processes.Mathematical Biosciences210, 490–507 (2007)

work page 2007
[54]

B., Diserud, O

Solbu, E. B., Diserud, O. H., K˚ al˚ as, J. A. & Engen, S. Heterogeneity among species and community dynamics—norwegian bird communities as a case study.Ecological Modelling388, 13–23 (2018)

work page 2018
[55]

Kendall, M. G. A new measure of rank correlation.Biometrika30, 81–93 (1938)

work page 1938
[56]

A.et al.Critical slowing down as early warning for the onset and termination of depression.Proceedings of the National Academy of Sciences 111, 87–92 (2014)

van de Leemput, I. A.et al.Critical slowing down as early warning for the onset and termination of depression.Proceedings of the National Academy of Sciences 111, 87–92 (2014)

work page 2014
[57]

M., Levin, S

May, R. M., Levin, S. A. & Sugihara, G. Ecology for bankers.Nature451, 893–894 (2008)

work page 2008
[58]

& Lambiotte, R

Neuh¨ auser, L., Mellor, A. & Lambiotte, R. Multibody interactions and nonlinear consensus dynamics on networked systems.Physical Review E101, 032310 (2020)

work page 2020
[59]

& Zhang, Y

Delabays, R., De Pasquale, G., D¨ orfler, F. & Zhang, Y. Hypergraph reconstruction from dynamics.Nature Communications16, 2691 (2025)

work page 2025
[60]

& Barab´ asi, A.-L

Barzel, B. & Barab´ asi, A.-L. Universality in network dynamics.Nature Physics9, 673–681 (2013). 25

work page 2013
[61]

& Barab´ asi, A.-L

Barzel, B., Liu, Y.-Y. & Barab´ asi, A.-L. Constructing minimal models for complex system dynamics.Nature Communications6, 7186 (2015)

work page 2015
[62]

& Yang, B

Hu, J., Cui, J. & Yang, B. Learning interpretable network dynamics via universal neural symbolic regression.Nature Communications16, 6226 (2025)

work page 2025
[63]

On explaining the surprising success of reservoir computing forecaster of chaos? the universal machine learning dynamical system with contrast to var and dmd.Chaos31, 013108 (2021)

Bollt, E. On explaining the surprising success of reservoir computing forecaster of chaos? the universal machine learning dynamical system with contrast to var and dmd.Chaos31, 013108 (2021)

work page 2021
[64]

Golub, G. H. & Van Loan, C. F.Matrix Computations(JHU press, 2013)

work page 2013
[65]

A new look at the statistical model identification.IEEE Transactions on Automatic Control19, 716–723 (1974)

Akaike, H. A new look at the statistical model identification.IEEE Transactions on Automatic Control19, 716–723 (1974)

work page 1974
[66]

Flores, B. E. A pragmatic view of accuracy measurement in forecasting.Omega 14, 93–98 (1986). 26

work page 1986