CoChat

Designing attractive models via automated identification of chaotic and oscillatory dynamical regimes

Open access
Silk, Daniel, Kirk, Paul D.W., Barnes, Chris P., Toni, Tina, Rose, Anna, Moon, Simon, Dallman, Margaret J., Stumpf, Michael P.H. · 2011 · Nature Communications

The DOI matches its Crossref record; 48 of the 48 references it cites were found in a registry or an index.

Abstract

Chaos and oscillations continue to capture the interest of both the scientific and public domains. Yet despite the importance of these qualitative features, most attempts at constructing mathematical models of such phenomena have taken an indirect, quantitative approach, for example, by fitting models to a finite number of data points. Here we develop a qualitative inference framework that allows us to both reverse-engineer and design systems exhibiting these and other dynamical behaviours by directly specifying the desired characteristics of the underlying dynamical attractor. This change in perspective from quantitative to qualitative dynamics, provides fundamental and new insights into the properties of dynamical systems. Modelling of chaos and oscillations is usually done indirectly and quantitatively by fitting models to a finite number of data-points. Here, a qualitative framework is developed where the characteristics of the underlying dynamical system are directly specified, revealing new properties of such systems.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1038/ncomms1496

Cites

48 of 48 references verified0 citation errors · 0 not checkable

Statistics

Figures

Figure 1 Encoding and inferring the desired dynamics. (a) LEs, λ0, ... λn, characterize the contraction/expansion of an initially small perturbation, ɛ0, to the system. (b) The leading LE determines the principal dynamics and characteristics of the attractor of a dynamical system. For λ0<0, the attractor will be a stable fixed point; stable oscillating solutions will be obtained, if λ0=0; for λ0>0 we observe chaos and the system will exhibit a so-called strange attractor; if more than one LE is positive, then we speak of hyperchaos and the attractor will exhibit behaviour with similar statistical properties to white noise. (c) Key steps in the UKF for qualitative inference. At the kth iteration, the current prior parameter distribution is formed by perturbing the previous posterior, θk, with the process noise vk. The distribution of LEs for the model f induced by the prior parameter distribution is calculated via the LE estimation routine L and the unscented transform. Comparing the mean LE, , to the target LE, λtarget, the prior parameters are updated using the UKF update equations. As the filter proceeds, parameters are found that locally minimise the sum of squared error between target and estimated LEs.
Figure 1 Encoding and inferring the desired dynamics. (a) LEs, λ0, ... λn, characterize the contraction/expansion of an initially small perturbation, ɛ0, to the system. (b) The leading LE determines the principal dynamics and characteristics of the attractor of a dynamical system. For λ0<0, the attractor will be a stable fixed point; stable oscillating solutions will be obtained, if λ0=0; for λ0>0 we observe chaos and the system will exhibit a so-called strange attractor; if more than one LE is positive, then we speak of hyperchaos and the attractor will exhibit behaviour with similar statistical properties to white noise. (c) Key steps in the UKF for qualitative inference. At the kth iteration, the current prior parameter distribution is formed by perturbing the previous posterior, θk, with the process noise vk. The distribution of LEs for the model f induced by the prior parameter distribution is calculated via the LE estimation routine L and the unscented transform. Comparing the mean LE, , to the target LE, λtarget, the prior parameters are updated using the UKF update equations. As the filter proceeds, parameters are found that locally minimise the sum of squared error between target and estimated LEs.CC BY-NC-ND · PubMed Central
Figure 2 Detecting and controlling chaos and oscillations. Plots show the estimated parameters at successive iterations of the UKF. Snapshots of the developing attractor are shown above the plots. The sum of squared error (E) is indicated for different sections of the parameter trajectories. (a) Diagram of the circuit investigated. (b) The filter is able to drive the electric circuit between oscillations and chaos in less than 10 iterations. (c) Parameter trajectories for a simple model of the Hes1 regulatory system that yield oscillations. Several regions in parameter space can be identified that exhibit oscillatory behaviour. (d) Examples of trajectories generated from a region in parameter space that was found using our approach. Here we used the qualitative inference procedure to elicit a prior to be used for parameter inference. Trajectories were sampled from the prior. Data are indicated by red circles and represent fold change in Hes1 mRNA; the blue strips indicate the confidence intervals obtained using Gaussian process regression, in which standard Gaussian noise is assumed, with maximum marginal likelihood estimates for the other hyperparameters21.
Figure 2 Detecting and controlling chaos and oscillations. Plots show the estimated parameters at successive iterations of the UKF. Snapshots of the developing attractor are shown above the plots. The sum of squared error (E) is indicated for different sections of the parameter trajectories. (a) Diagram of the circuit investigated. (b) The filter is able to drive the electric circuit between oscillations and chaos in less than 10 iterations. (c) Parameter trajectories for a simple model of the Hes1 regulatory system that yield oscillations. Several regions in parameter space can be identified that exhibit oscillatory behaviour. (d) Examples of trajectories generated from a region in parameter space that was found using our approach. Here we used the qualitative inference procedure to elicit a prior to be used for parameter inference. Trajectories were sampled from the prior. Data are indicated by red circles and represent fold change in Hes1 mRNA; the blue strips indicate the confidence intervals obtained using Gaussian process regression, in which standard Gaussian noise is assumed, with maximum marginal likelihood estimates for the other hyperparameters21.CC BY-NC-ND · PubMed Central
Figure 3 Designing attractive models. (a) Inferring a complete spectrum. After only 22 iterations, the characteristic 'butterfly' strange attractor emerges. The final parameters and LEs are σ=10.2, ρ=29.2, β=2.45 and (0.899, 2.74×10−4, −14.6). (b) A function of the LEs, the Kaplan–Yorke fractal dimension may also be used to specify the desired attractor. Here parameters for a target dimension of 1 are found for the Lorenz system within 20 iterations, giving rise to a limit cycle as required by the theory. (c) Three-dimensional projections of the hyperchaotic system with parameter vector (a, b, c, d, e, f)=(49.98, 35.86, 30.5, 1.35, 36.6, 33.8) and corresponding LEs (31.8, 16.8, −19.1, −71.4). Within a few iterations, our algorithm was able to drive the system towards an attractor characterized by LEs twice the size of any that had previously been reported.
Figure 3 Designing attractive models. (a) Inferring a complete spectrum. After only 22 iterations, the characteristic 'butterfly' strange attractor emerges. The final parameters and LEs are σ=10.2, ρ=29.2, β=2.45 and (0.899, 2.74×10−4, −14.6). (b) A function of the LEs, the Kaplan–Yorke fractal dimension may also be used to specify the desired attractor. Here parameters for a target dimension of 1 are found for the Lorenz system within 20 iterations, giving rise to a limit cycle as required by the theory. (c) Three-dimensional projections of the hyperchaotic system with parameter vector (a, b, c, d, e, f)=(49.98, 35.86, 30.5, 1.35, 36.6, 33.8) and corresponding LEs (31.8, 16.8, −19.1, −71.4). Within a few iterations, our algorithm was able to drive the system towards an attractor characterized by LEs twice the size of any that had previously been reported.CC BY-NC-ND · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 48 references

Research works printed but not linked: 1

  1. Julier, S. & Uhlmann, J. A general method for approximating nonlinear transformations of probability distributions. Tech. Rep. RRG, Dept. of Engineering Science, University of Oxford (1996).

Cited by

See all 76

Metadata

Authors
Silk, D., Kirk, P. D. W., Barnes, C. P., Toni, T., Rose, A., Moon, S., Dallman, M. J., & Stumpf, M. P. H.
Affiliations
Imperial College London (GB)
Corresponding author
Stumpf, Michael P.H.
Venue
Nature Communications
Published
2011-10-04
Volume
2
Issue
1
Pages
489-489
ISSN
2041-1723 (ISSN-L)
OpenAlex
W2059659399
Cited by
76
References
48
Citations by year
2026: 2 · 2025: 3 · 2024: 2 · 2023: 9 · 2022: 6 · 2021: 9 · 2020: 2 · 2019: 7 · 2018: 4 · 2017: 3 · 2016: 3 · 2015: 5 · 2014: 7 · 2013: 6 · 2012: 7
FWCI
5.13
Type
journal article
Language
en
Open access
Licence
Locations
  • Nature Communications (publishedVersion, cc-by-nc-nd)
  • PubMed Central (publishedVersion, cc-by-nc-nd)
  • PubMed (publishedVersion)
  • arXiv (Cornell University) (submittedVersion)
  • Minerva Access (University of Melbourne) (submittedVersion, cc-by-nc-nd)
  • RePEc: Research Papers in Economics (submittedVersion)
  • UCL Discovery (University College London) (submittedVersion)
  • arXiv (Cornell University)
Funders
  • Wellcome Trust
  • Imperial College London
  • Directorate for Biological Sciences
  • Biotechnology and Biological Sciences Research Council (BB/F005210/1, BB/G007934/1, BB/FO05210/1)
Topics
  • Physical Sciences › Physics and Astronomy › Statistical and Nonlinear Physics › Chaos control and synchronization
  • Physical Sciences › Mathematics › Mathematical Physics › Mathematical Dynamics and Fractals
  • Physical Sciences › Physics and Astronomy › Statistical and Nonlinear Physics › Model Reduction and Neural Networks
Keywords
chaotic dynamics, oscillatory dynamics, dynamical attractors, model design, dynamical regimes, qualitative dynamics
MeSH terms
Automation, Models, Theoretical, Signal Transduction, Nonlinear Dynamics

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Designing attractive models via automated identification of chaotic and oscillatory dynamical regimes. https://app.cochat.ai/papers/10.1038/ncomms1496
The paper
Silk, D., Kirk, P. D. W., Barnes, C. P., Toni, T., Rose, A., Moon, S., Dallman, M. J., & Stumpf, M. P. H. (2011). Designing attractive models via automated identification of chaotic and oscillatory dynamical regimes. Nature Communications. https://doi.org/10.1038/ncomms1496

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.