CoChat

Generalised Filtering

Open access
Friston, Karl, Stephan, Klaas, Li, Baojuan, Daunizeau, Jean · 2010 · Mathematical Problems in Engineering

The DOI matches its Crossref record; 28 of the 28 references it cites were found in a registry or an index.

Abstract

We describe a Bayesian filtering scheme for nonlinear state‐space models in continuous time. This scheme is called Generalised Filtering and furnishes posterior (conditional) densities on hidden states and unknown parameters generating observed data. Crucially, the scheme operates online, assimilating data to optimize the conditional density on time‐varying states and time‐invariant parameters. In contrast to Kalman and Particle smoothing, Generalised Filtering does not require a backwards pass. In contrast to variational schemes, it does not assume conditional independence between the states and parameters. Generalised Filtering optimises the conditional density with respect to a free‐energy bound on the model′s log‐evidence. This optimisation uses the generalised motion of hidden states and parameters, under the prior assumption that the motion of the parameters is small. We describe the scheme, present comparative evaluations with a fixed‐form variational version, and conclude with an illustrative application to a nonlinear state‐space model of brain imaging time‐series.

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 · 10.1155/2010/621670

Cites

28 of 28 references verified0 citation errors · 0 not checkable

Statistics

Figures

Figure 1 This graph shows the kernels implied by the recognition dynamics in parenleftmath 2.9 parenrightmath that accumulate evidence for changes in the conditional estimates of model parameters. These kernels are applied to the history of free-energy gradients to produce changes in the conditional mean parenleftmath solid line parenrightmath and its motion parenleftmath broken line parenrightmath . The kernels are derived from a standard Volterra-series expansion about the true conditional mean, where, in this example, κ equalmath 4.
Figure 1 This graph shows the kernels implied by the recognition dynamics in parenleftmath 2.9 parenrightmath that accumulate evidence for changes in the conditional estimates of model parameters. These kernels are applied to the history of free-energy gradients to produce changes in the conditional mean parenleftmath solid line parenrightmath and its motion parenleftmath broken line parenrightmath . The kernels are derived from a standard Volterra-series expansion about the true conditional mean, where, in this example, κ equalmath 4.CC-BY · From the PDF
Figure 2 The linear state-space model and an example of the data it generates: the upper left panel shows simulated data in terms of the output due to hidden states parenleftmath coloured lines parenrightmath and observation noise parenleftmath red lines parenrightmath . The parenleftmath noisy parenrightmath dynamics of the hidden states are shown in the upper right panels parenleftmath blue lines parenrightmath , which are the response to the cause or input on the lower left. The generative model is shown as a Bayesian dependency graph on the lower right.
Figure 2 The linear state-space model and an example of the data it generates: the upper left panel shows simulated data in terms of the output due to hidden states parenleftmath coloured lines parenrightmath and observation noise parenleftmath red lines parenrightmath . The parenleftmath noisy parenrightmath dynamics of the hidden states are shown in the upper right panels parenleftmath blue lines parenrightmath , which are the response to the cause or input on the lower left. The generative model is shown as a Bayesian dependency graph on the lower right.CC-BY · From the PDF
Figure 3 Conditional estimates during the first iteration of Generalised Filtering. This format will be used in subsequent figures and summarizes the predictions and conditional densities on the states of a hierarchical dynamic model. The first parenleftmath upper left parenrightmath panel shows the predicted response parenleftmath coloured lines parenrightmath and the error parenleftmath red lines parenrightmath on this response parenleftmath their sum corresponds to observed data parenrightmath . For the hidden states parenleftmath upper right parenrightmath and causes parenleftmath lower left parenrightmath the conditional mode is depicted by a blue line and the 90% conditional confidence intervals parenleftmath regions parenrightmath by the grey area. The lower right panels show the optimisation of the conditional means of the free parameters parenleftmath above parenrightmath and log-precisions parenleftmath below parenrightmath as a function of time.
Figure 3 Conditional estimates during the first iteration of Generalised Filtering. This format will be used in subsequent figures and summarizes the predictions and conditional densities on the states of a hierarchical dynamic model. The first parenleftmath upper left parenrightmath panel shows the predicted response parenleftmath coloured lines parenrightmath and the error parenleftmath red lines parenrightmath on this response parenleftmath their sum corresponds to observed data parenrightmath . For the hidden states parenleftmath upper right parenrightmath and causes parenleftmath lower left parenrightmath the conditional mode is depicted by a blue line and the 90% conditional confidence intervals parenleftmath regions parenrightmath by the grey area. The lower right panels show the optimisation of the conditional means of the free parameters parenleftmath above parenrightmath and log-precisions parenleftmath below parenrightmath as a function of time.CC-BY · From the PDF
Figure 4 Conditional estimates after convergence of the GF scheme. This figure uses the same format as the previous figure but includes the true values of states used to generate the response parenleftmath broken grey lines parenrightmath . Furthermore, time-dependent estimates of the parameters and precisions have been replaced with the conditional moments of the Bayesian average of the parameters over time. The black bars are the true values used to generate the data and the white bars are the conditional means. 90% confidence intervals are shown in red.
Figure 4 Conditional estimates after convergence of the GF scheme. This figure uses the same format as the previous figure but includes the true values of states used to generate the response parenleftmath broken grey lines parenrightmath . Furthermore, time-dependent estimates of the parameters and precisions have been replaced with the conditional moments of the Bayesian average of the parameters over time. The black bars are the true values used to generate the data and the white bars are the conditional means. 90% confidence intervals are shown in red.CC-BY · From the PDF
Figure 5 Conditional estimates after convergence of the DEM scheme. This is exactly the same as the previous figure but shows the conditional estimates following Dynamic Expectation Maximisation of the data in Figure 2. The key difference is manifested in terms of smaller confidence tubes and intervals on the states and parameters, respectively, which are overconfident in relation to GF parenleftmath cf. Figure 4 parenrightmath .
Figure 5 Conditional estimates after convergence of the DEM scheme. This is exactly the same as the previous figure but shows the conditional estimates following Dynamic Expectation Maximisation of the data in Figure 2. The key difference is manifested in terms of smaller confidence tubes and intervals on the states and parameters, respectively, which are overconfident in relation to GF parenleftmath cf. Figure 4 parenrightmath .CC-BY · From the PDF
Figure 6 Comparison of Generalised Filtering and DEM. Upper left: negative free-action for GF parenleftmath solid line parenrightmath and DEM parenleftmath dotted line parenrightmath as a function of iteration number. Upper right: conditional moments of the logprecisions, shown for GF parenleftmath grey bars parenrightmath and DEM parenleftmath white bars parenrightmath , in relation to the true values used to generate the data parenleftmath black bars parenrightmath . 90% confidence intervals are shown in red. The lower panels show the conditional moments of the parameters parenleftmath left parenrightmath and log-precisions parenleftmath right parenrightmath as a function of time, after convergence of the GF scheme. The conductional means parenleftmath minus the prior expectation of the parameters parenrightmath are shown as blue lines within their 90% confidence tubes parenleftmath grey regions parenrightmath .
Figure 6 Comparison of Generalised Filtering and DEM. Upper left: negative free-action for GF parenleftmath solid line parenrightmath and DEM parenleftmath dotted line parenrightmath as a function of iteration number. Upper right: conditional moments of the logprecisions, shown for GF parenleftmath grey bars parenrightmath and DEM parenleftmath white bars parenrightmath , in relation to the true values used to generate the data parenleftmath black bars parenrightmath . 90% confidence intervals are shown in red. The lower panels show the conditional moments of the parameters parenleftmath left parenrightmath and log-precisions parenleftmath right parenrightmath as a function of time, after convergence of the GF scheme. The conductional means parenleftmath minus the prior expectation of the parameters parenrightmath are shown as blue lines within their 90% confidence tubes parenleftmath grey regions parenrightmath .CC-BY · From the PDF
Figure 7 Conditional estimates following Generalised Filtering of the empirical brain imaging time-series. This figure adopts the same format as Figure 4; however, only three parenleftmath of the eight parenrightmath parameter estimates are shown parenleftmath lower right parenrightmath . These are the conditional means parenleftmath grey bars parenrightmath and 90% confidence intervals parenleftmath red bars parenrightmath of coupling parameters that mediate the effects of vision, motion, and attention on neuronal activity.
Figure 7 Conditional estimates following Generalised Filtering of the empirical brain imaging time-series. This figure adopts the same format as Figure 4; however, only three parenleftmath of the eight parenrightmath parameter estimates are shown parenleftmath lower right parenrightmath . These are the conditional means parenleftmath grey bars parenrightmath and 90% confidence intervals parenleftmath red bars parenrightmath of coupling parameters that mediate the effects of vision, motion, and attention on neuronal activity.CC-BY · From the PDF
Figure 8 Parameter and precision estimates from the analysis of the empirical data presented in the previous figure. Upper left: conditional means parenleftmath grey bars parenrightmath and 90% confidence intervals parenleftmath red bars parenrightmath of all parenleftmath eight parenrightmath free parameters in this nonlinear model. Upper right: the corresponding conditional covariances are shown in image format parenleftmath with arbitrary scaling parenrightmath . The lower panels show the time-dependent changes in conditional moments as a function of scan number for the parameters parenleftmath minus their prior expectation; left parenrightmath and the log-precisions parenleftmath right parenrightmath . We have focused on the precision of the first parenleftmath vision parenrightmath coupling parameter parenleftmath grey area parenrightmath in this figure.
Figure 8 Parameter and precision estimates from the analysis of the empirical data presented in the previous figure. Upper left: conditional means parenleftmath grey bars parenrightmath and 90% confidence intervals parenleftmath red bars parenrightmath of all parenleftmath eight parenrightmath free parameters in this nonlinear model. Upper right: the corresponding conditional covariances are shown in image format parenleftmath with arbitrary scaling parenrightmath . The lower panels show the time-dependent changes in conditional moments as a function of scan number for the parameters parenleftmath minus their prior expectation; left parenrightmath and the log-precisions parenleftmath right parenrightmath . We have focused on the precision of the first parenleftmath vision parenrightmath coupling parameter parenleftmath grey area parenrightmath in this figure.CC-BY · From the PDF
Figure 9 These are the same results shown in Figure 7 but focussing on the conditional expectations of the hidden states and neuronal causes over the first 128 parenleftmath 3.22 second parenrightmath time bins. Left panel: the hidden states are overlaid on periods parenleftmath grey bars parenrightmath of visual motion. These hidden states correspond to flowinducing signal, flow, volume, and deoxyhaemoglobin parenleftmath dHb parenrightmath . It can be seen that neuronal activity, shown in the right panel, induces a transient burst of signal parenleftmath blue parenrightmath , which is rapidly suppressed by the resulting increase in flow parenleftmath green parenrightmath . The increase in flow dilates the venous capillary bed to increase volume parenleftmath red parenrightmath and dilute deoxyhaemoglobin parenleftmath cyan parenrightmath . The concentration of deoxyhaemoglobin parenleftmath involving volume and dHb parenrightmath determines the measured response.
Figure 9 These are the same results shown in Figure 7 but focussing on the conditional expectations of the hidden states and neuronal causes over the first 128 parenleftmath 3.22 second parenrightmath time bins. Left panel: the hidden states are overlaid on periods parenleftmath grey bars parenrightmath of visual motion. These hidden states correspond to flowinducing signal, flow, volume, and deoxyhaemoglobin parenleftmath dHb parenrightmath . It can be seen that neuronal activity, shown in the right panel, induces a transient burst of signal parenleftmath blue parenrightmath , which is rapidly suppressed by the resulting increase in flow parenleftmath green parenrightmath . The increase in flow dilates the venous capillary bed to increase volume parenleftmath red parenrightmath and dilute deoxyhaemoglobin parenleftmath cyan parenrightmath . The concentration of deoxyhaemoglobin parenleftmath involving volume and dHb parenrightmath determines the measured response.CC-BY · From the PDF
Figure 10 The equivalent results for the hemodynamic deconvolution using DEM . These densities should be compared with those in Figure 7 that were obtained using Generalised Filtering.
Figure 10 The equivalent results for the hemodynamic deconvolution using DEM . These densities should be compared with those in Figure 7 that were obtained using Generalised Filtering.CC-BY · From the PDF
Image 11
CC-BY · From the PDF
Image 12
CC-BY · From the PDF
Image 13
CC-BY · From the PDF
Image 14
CC-BY · From the PDF
Image 15
CC-BY · From the PDF
Image 16
CC-BY · From the PDF
Image 17
CC-BY · From the PDF
Image 18
CC-BY · From the PDF
Image 19
CC-BY · From the PDF
Image 20
CC-BY · From the PDF
Image 21
CC-BY · From the PDF
Image 22
CC-BY · From the PDF
Image 23
CC-BY · From the PDF
Image 24
CC-BY · From the PDF
Image 25
CC-BY · From the PDF
Image 26
CC-BY · From the PDF
Image 27
CC-BY · From the PDF
Image 28
CC-BY · From the PDF
Image 29
CC-BY · From the PDF

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

Research works printed but not linked: 3

  1. C. Archambeau, D. Cornford, M. Opper, and J. Shawe-Taylor, "Gaussian process approximations of stochastic differential equations," in Proceedings of the Journal of Machine Learning Research Workshop and Conference, vol. 1, pp. 1-16, October 2007.
  2. M. J. Beal and Z. Ghahramani, "The variational Bayesian EM algorithm for incomplete data: with application to scoring graphical model structures," in Bayesian Statistics, J. M. Bernardo, M. J. Bayarri, J. O. Berger, et al., Eds., chapter 7, Oxford University Press, Oxford, UK, 2003.
  3. T. Ozaki, "A bridge between nonlinear time series models and nonlinear stochastic dynamical systems: a local linearization approach," Statistica Sinica, vol. 2, no. 1, pp. 113-135, 1992.

Cited by

See all 171

Metadata

Authors
Friston, K., Stephan, K., Li, B., & Daunizeau, J.
Affiliations
Wellcome Centre for Human Neuroimaging (GB)
Corresponding author
Friston, Karl
Venue
Mathematical Problems in Engineering
Publisher
Wiley
Published
2010-01
Volume
2010
Issue
1
ISSN
  • 1024-123X (ISSN-L)
  • 1563-5147
OpenAlex
W4249210107
Cited by
171
References
25
Citations by year
2026: 8 · 2025: 14 · 2024: 6 · 2023: 11 · 2022: 11 · 2021: 10 · 2020: 7 · 2019: 11 · 2018: 9 · 2017: 13 · 2016: 10 · 2015: 19 · 2014: 10 · 2013: 8 · 2012: 13
FWCI
10.83
Type
journal article
Language
en
Open access
Locations
  • Mathematical Problems in Engineering (publishedVersion, cc-by)
  • RePEc: Research Papers in Economics (submittedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
  • UCL Discovery (University College London) (submittedVersion)
Funders
China Scholarship Council
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › Target Tracking and Data Fusion in Sensor Networks
  • Physical Sciences › Computer Science › Artificial Intelligence › Gaussian Processes and Bayesian Inference
  • Physical Sciences › Engineering › Control and Systems Engineering › Control Systems and Identification
Keywords
Bayesian filtering, nonlinear state-space models, continuous-time models, internet filtering, posterior density estimation, free energy bounds, variational inference, parameter estimation

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

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.