Approximate Bayesian computation (ABC) gives exact results under the assumption of model error
Open accessThe DOI matches its Crossref record; 40 of the 41 references it cites were found in a registry or an index.
Abstract
Approximate Bayesian computation (ABC) or likelihood-free inference algorithms are used to find approximations to posterior distributions without making explicit use of the likelihood function, depending instead on simulation of sample data sets from the model. In this paper we show that under the assumption of the existence of a uniform additive model error term, ABC algorithms give exact results when sufficient summaries are used. This interpretation allows the approximation made in many previous application papers to be understood, and should guide the choice of metric and tolerance in future work. ABC algorithms can be generalized by replacing the 0-1 cut-off with an acceptance probability that varies with the distance of the simulated data from the observed data. The acceptance density gives the distribution of the error term, enabling the uniform error usually used to be replaced by a general distribution. This generalization can also be applied to approximate Markov chain Monte Carlo algorithms. In light of this work, ABC algorithms can be seen as calibration techniques for implicit stochastic models, inferring parameter values in light of the computer model, data, prior beliefs about the parameter values, and any measurement or model errors.
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.
Integrity signals
Registry
Cites
Statistics
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.
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?
In the citing paper
“The key strength of ABC approaches is that they can be applied to problems with intractable likelihoods35,36 (for example, complex stochastic models).”
From A framework for parameter estimation and model selection from experimental data in systems biology using approximate Bayesian computation (2014) · cites this paper as [35]
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
Kass, R. E., & Raftery, A. E. (1995) · Journal of the American Statistical Association · doi:10.1080/01621459.1995.10476572 · Cited by 12,558
Approximate Bayesian computation in population genetics
Beaumont, M. A., Zhang, W., & Balding, D. J. (2002) · Genetics · doi:10.1093/genetics/162.4.2025 · Cited by 2,932
Monte Carlo Strategies in Scientific Computing
Liu, J. S. (2004) · Springer Series in Statistics · doi:10.1007/978-0-387-76371-2 · Cited by 2,690
Toni, T., Welch, D., Strelkowa, N., Ipsen, A., & Stumpf, M. P. H. (2009) · Journal of The Royal Society Interface · doi:10.1098/rsif.2008.0172 · Cited by 1,498
Monte Carlo Strategies in Scientific Computing
Hesterberg, T. (2002) · Technometrics · doi:10.1198/tech.2002.s85 · Cited by 1,409
Markov chain Monte Carlo without likelihoods
Marjoram, P., Molitor, J., Plagnol, V., & Tavaré, S. (2003) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.0306899100 · Cited by 1,199
Population growth of human Y chromosomes: a study of Y chromosome microsatellites
Pritchard, J. K., Seielstad, M. T., Perez-Lezaun, A., & Feldman, M. W. (1999) · Molecular Biology and Evolution · doi:10.1093/oxfordjournals.molbev.a026091 · Cited by 1,080
Computer Model Calibration Using High-Dimensional Output
Higdon, D., Gattiker, J., Williams, B., & Rightley, M. (2008) · Journal of the American Statistical Association · doi:10.1198/016214507000000888 · Cited by 867
Sequential Monte Carlo without likelihoods
Sisson, S. A., Fan, Y., & Tanaka, M. M. (2007) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.0607208104 · Cited by 856
Cornuet, J.-M., Santos, F., Beaumont, M. A., Robert, C. P., Marin, J.-M., Balding, D. J., Guillemaud, T., & Estoup, A. (2008) · Bioinformatics · doi:10.1093/bioinformatics/btn514 · Cited by 690
Research works printed but not linked: 33
- Wilkinson, 2007, Bayesian inference of primate divergence times Ph thesis University of
- Nunes, M. A., Balding, D. J., 2010. On optimal selection of summary statistics for approximate Bayesian computation. Stat. Appl. Genet. Mo. B. 9, Article 34. · DOI · Paper page
- Blum, M., Tran, V., 2010. HIV with contact tracing: a case study in approximate Bayesian computation. Biostatistics 11 (4), 644-660. · DOI
- Shibata, 2008, Siegmund Marjoram Modeling DNA methylation in apopulation of cancer cells article, Stat Appl Genet Mo B, 18
- Plagnol, 2004, Approximate Bayesian computation and In Proceedings of Monte Carlo and Quasi - Monte Carlo Verlag, Methods, 99
- Ratmann, 2007, Using likelihood - free inference to compare evolutionary dynamics of the protein networks ofH pyloriandP falciparum, Comput Biol, 2266
- Fearnhead, P., Prangle, D., 2012. Constructing summary statistics for approximate Bayesian computation: semi-automatic approximate Bayesian computation. J. Roy. Stat. Soc. Ser. B 74, 419-474. · DOI · Paper page
- Kennedy, M., O'Hagan, A., 2001. Bayesian calibration of computer models (with discussion). J. R. Statist. Soc. Ser. B 63, 425-464.
- Robert, C. P., Cornuet, J. M., Marin, J. M., Pillai, N. S., 2011. Lack of confidence in approximate Bayesian computation model choice. P. Natl. Acad. Sci. USA 108, 15112-15117.
- Toni, T., Welch, D., Strelkowa, N., Ipsen, A., Stumpf, M. P. H., 2009. Approximate Bayesian Computation scheme for parameter inference and model selection in dynamical systems. J. R. Soc. Interface 6 (31), 187-202. · DOI · Paper page
- Campbell, 2006, Statistical calibration of computer simulations, Eng Syst Safe, 1358, 10.1016/j.ress.2005.11.032 · DOI
- Sunnaker, M., Busetto, A. G., Numminen, E., Corander, J., Foll, M., Dessimoz, C., 2013. Approximate bayesian computation. PLoS Comput Biol 9, e1002803.
- Cornuet, 2008, Inferring population history with DIY ABC a user - friendly approach to approximate Bayesian computation, Bioinformatics, 24, 2713, 10.1093/bioinformatics/btn514 · DOI · Paper page
- Beaumont, 2002, Balding Approximate Bayesian computatation in population genetics, Genetics, 162
- Barnes, 2012, Considerate approaches to constructing summary statistics for ABC model selection, Stat Comput, 22, 1181, 10.1007/s11222-012-9335-7 · DOI · Paper page
- Murray, I., Ghahramani, Z., MacKay, D. J. C., 2006. MCMC for doublyintractable distributions. In: Proceedings of the 22nd Annual Conference on Uncertainty in Artificial Intelligence. UAI.
- Marjoram, P., Molitor, J., Plagnol, V ., Tavar´ e, S., 2003. Markov chain Monte Carlo without likelihoods. P. Natl. Acad. Sci. USA 100 (26), 15324-15328.
- Pritchard, J. K., Seielstad, M. T., Perez-Lezaun, A., Feldman, M. W., 1999. Population growth of human Y chromosomes: a study of Y chromosome microsatellites. Mol. Biol. Evol. 16 (12), 1791-1798. · DOI · Paper page
- Goldstein, 2009, Bayesian modelling and inference for physical systems, Stat Plan Infer, 139
- Sisson, S. A., Fan, Y., Tanaka, M. M., 2007. Sequential Monte Carlo without likelihoods. Proc. Natl. Acad. Sci. USA 104 (6), 1760-1765. · DOI · Paper page
- Wilkinson, R. D., 2007. Bayesian inference of primate divergence times. Ph.D. thesis, University of Cambridge.
- Higdon, 2008, model calibration using high - dimensional output, Computer Am Assoc, 103
- Peters, 2012, On sequential Monte Carlo partial rejection control and approximate Bayesian computation, Stat Comput, 22, 1209, 10.1007/s11222-012-9315-y · DOI · Paper page
- Hamilton, 2005, Bayesian estimation of recent migration rates after a spatial expansion, Genetics, 170
- Doucet, 2012, Del Moral An adaptive sequential Monte Carlo method for approximate Bayesian computation, Stat Comput, 22, 1009, 10.1007/s11222-011-9271-y · DOI · Paper page
- Foll, 2008, An approximate Bayesian computation approach to overcome biases that arise when using amplifed fragment length polymorphism markers to study population structure, Genetics, 179
- Kass, 1995, factors, Am Assoc, 773, 10.1080/01621459.1995.10476572 · DOI · Paper page
- Joyce, 2008, Marjoram Approximately sufficient Bayesian computation article, statistics Stat Appl Genet Mo B, 26
- Kennedy, 2001, Bayesian calibration of computer models with discussion, Statist, 425
- Tanaka, M. M., Francis, A. R., Luciani, F., Sisson, S. A., 2006. Using approximate Bayesian computation to estimate tuberculosis transmission parameters from genotype data. Genetics 173, 1511-1520.
- Beaumont, 2009, for ABC algorithms the ABC, Biometrika, 983, 10.1093/biomet/asp052 · DOI · Paper page
- Wilkinson, R. D., Tavar´ e, S., 2009. Estimating primate divergence times by using conditioned birth-and-death processes. Theor. Popul. Biol. 75, 278-285. · DOI
- Drovandi, 2011, Faddy Approximate Bayesian computation using indirect inference, Stat Soc, 317
Cited by
Items 101–200 of 334
A rare event approach to high-dimensional approximate Bayesian computation
Prangle, D., Everitt, R. G., & Kypraios, T. (2017) · Statistics and Computing · doi:10.1007/s11222-017-9764-4 · Cited by 29
“Alternative kernel functions have also been used in ABC (e.g., Wilkinson 2013) such as a Gaussian: k ( y ; (cid:4)) ∝ exp [− (cid:4) 2 d ( y , y obs ) ] .”
MethodA practical guide to pseudo-marginal methods for computational inference in systems biology
Warne, D. J., Baker, R. E., & Simpson, M. J. (2020) · Journal of Theoretical Biology · doi:10.1016/j.jtbi.2020.110255 · Cited by 29
“In general, one requires the summary statistics to be sufficient statistics (Fearnhead and Prangle, 2012) and (cid:15) to be of similar order to the observation error (Toni et al., 2009; Wilkinson, 2013) to obtain accurate posteriors for the purposes of inference.”
ResultModularized Bayesian analyses and cutting feedback in likelihood-free inference
Chakraborty, A., Nott, D. J., Drovandi, C., Frazier, D. T., & Sisson, S. A. (2023) · Statistics and Computing · doi:10.1007/s11222-023-10207-5 · Cited by 28
“In traditional ABC methods the ABC kernel can be interpreted in terms of an allowance for model misspecification (Wilkinson 2013), and an interesting avenue for future research might be to pursue ABC approaches to the semi-modular inference framework making use of this interpretation as a way of partially cutting feed…”
MethodGeneralized Bayesian Inference for Scientific Simulators via Amortized Cost Estimation
Gao, R., Deistler, M., & Macke, J. H. (2023) · neural information processing systems · doi:10.52202/075280-3514 · Cited by 28
Drovandi, C., Pettitt, A. N., & McCutchan, R. A. (2015) · Bayesian Analysis · doi:10.1214/15-ba950 · Cited by 27
Misspecification-robust Sequential Neural Likelihood for Simulation-based Inference
Kelly, R., Nott, D. J., Frazier, D. T., Warne, D. J., & Drovandi, C. (2023) · Transactions on machine learning research · doi:10.48550/arxiv.2301.13368 · Cited by 27
“Although ABC is, in a certain sense, robust to misspecification, and BSL has been extended to handle incompatibility, they both remain inefficient in terms of the number of model simulations required.”
BackgroundMarin, J., Pudlo, P., Estoup, A., & Robert, C. P. (2018) · Chapman & Hall/CRC eBooks · doi:10.1201/9781315117195-6 · Cited by 26
“ows to innity, but we also benet from new results that set actual ABC implementations, with their nite computing power and strictly positive tolerances, within the range of other types of inference [38,39,40]. ABC now stands as an inference method that is justiable on its own ground. This approach may be the only solu…”
BackgroundBayesian Probabilistic Power Flow Analysis Using Jacobian Approximate Bayesian Computation
Zuluaga, C. D., & Álvarez, M. A. (2018) · IEEE Transactions on Power Systems · doi:10.1109/tpwrs.2018.2810641 · Cited by 26
“However, the algorithm would need to perform a large amount of simulations for accepting any sample [11].”
BackgroundMethodInfluentialGeneralized Posteriors in Approximate Bayesian ComputationPreprint
Schmon, S. M., Cannon, P. W., & Knoblauch, J. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2011.08644 · Cited by 26
“While this argument for interpreting the usual abc kernels as error distributions was previously made by Wilkinson (2013), it seems implausible that the correct measurement error of a complex simulator can be expressed in terms of a difference of low-dimensional summary statistics via Kh.”
BackgroundReverse Engineering Under Uncertainty
Kirk, P., Silk, D., & Stumpf, M. P. H. (2015) · Studies in mechanobiology, tissue engineering and biomaterials · doi:10.1007/978-3-319-21296-8_2 · Cited by 25
“For some problems it may not even be possible to write down the likelihood, in which case approximate likelihood techniques and approaches such as approximate Bayesian computation might be appropriate [52, 60].”
MethodA large-scale stochastic spatiotemporal model for Aedes albopictus-borne chikungunya epidemiology
Erguler, K., Chandra, N. L., Proestos, Y., Lelieveld, J., Christophides, G. K., & Parham, P. E. (2017) · PLoS ONE · doi:10.1371/journal.pone.0174293 · Cited by 24
Asymptotically exact inference in differentiable generative models
Graham, M. M., & Storkey, A. (2017) · Electronic Journal of Statistics · doi:10.1214/17-ejs1340si · Cited by 24
“Alternatively the kernel 𝑘 𝜖 can be given a modelling interpretation as representing uncertainty introduced by noise in the observations or mismatch between the unknown generative process by which the observed data was produced and the generative model [30, 37].”
BackgroundEtude de stratégies de gestion en temps réel pour des bâtiments énergétiquement performants
Robillart, M. (2015) · theses.fr (ABES) · doi:10.70675/8180d98dz6687z48f0zbb3eza3482dca435b · Cited by 24
Efficient Probabilistic Inference in the Quest for Physics Beyond the Standard Model
Baydin, A. G., Heinrich, L., Bhimji, W., Shao, L., Naderiparizi, S., Munk, A., Liu, J., Gram-Hansen, B., Louppe, G., Meadows, L., et al. (2019) · Oxford University Research Archive (ORA) (University of Oxford) · Cited by 23
Daly, A. C., Cooper, J., Gavaghan, D. J., & Holmes, C. C. (2017) · Journal of The Royal Society Interface · doi:10.1098/rsif.2017.0340 · Cited by 23
“Wilkinson suggests choosing for cεt the modal value of the acceptance kernel, where SE(θ) = 0 [44], which would improve the overall acceptance rate but still bound the function between 0 and 1.”
BackgroundMethodGenetic Load and Potential Mutational Meltdown in Cancer Cell Populations
Zhang, Y., Li, Y., Shen, X., Zhu, T., Tao, Y., Tao, Y., Li, X., Wang, D., Ma, Q., Hu, Z., et al. (2018) · Molecular Biology and Evolution · doi:10.1093/molbev/msy231 · Cited by 23
Improved calibration of building models using approximate Bayesian calibration and neural networks
Cant, K., & Evins, R. (2022) · Journal of Building Performance Simulation · doi:10.1080/19401493.2022.2137236 · Cited by 22
“The most basic ABC algorithm is the ABC-Rejection algorithm (Toni and Stumpf 2009).”
BackgroundMethodInfluentialMaclaren, O. J., Parker, A., Pin, C., Carding, S. R., Watson, A. J. M., Fletcher, A. G., Byrne, H. M., & Maini, P. K. (2017) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1005688 · Cited by 22
Distilling Importance Sampling for Likelihood Free Inference
Prangle, D., & Viscardi, C. (2023) · Journal of Computational and Graphical Statistics · doi:10.1080/10618600.2023.2175688 · Cited by 22
“Densities of the form (6) are often used in ABC and correspond to the posterior for data observed with independent N ( 0, (cid:2) 2 ) additive errors (Wilkinson 2013).”
MethodConstraining Cognitive Abstractions Through Bayesian Modeling
Turner, B. M. (2015) · Springer eBooks · doi:10.1007/978-1-4939-2236-9_10 · Cited by 22
Molloy, M. F., Galdo, M., Bahg, G., Liu, Q., & Turner, B. M. (2018) · Decision · doi:10.1037/dec0000097 · Cited by 21
“…methods require that observed and simulated data are compared by a set of summary statistics that adequately describe the distribution of both data sets (e.g., Toni, Welch, Strelkowa, Ipsen, & Stumpf, 2009; Turner & Sederberg, 2012; Turner & Van Zandt, 2012; Wilkinson, 2008; Wood, 2010).”
MethodABC methods for model choice in Gibbs random fields
Grelaud, A., Robert, C. P., & Marin, J.-M. (2009) · Comptes Rendus. Mathématique · doi:10.1016/j.crma.2008.12.009 · Cited by 20
Irvine, M. A., & Hollingsworth, T. D. (2018) · Epidemics · doi:10.1016/j.epidem.2018.05.009 · Cited by 20
“This particle θ is then accepted if this discrepancy is below a pre-de fi ned tolerance ϵ and rejected otherwise (Wilkinson, 2013) (e.g. for its fi rst use see Pritchard et al., 1999, and see Blum and Tran (2010) for a smoothed rejection scheme applied to fi tting an SIR model).”
MethodGuided Sequential ABC Schemes for Intractable Bayesian Models
Picchini, U., & Tamborrino, M. (2024) · Bayesian Analysis · doi:10.1214/24-ba1451 · Cited by 20
“A version of ABC-rejection, as given in Wilkinson [2013b], proposes parameter θ ∗ yielding a distance || s ∗ − s y || .”
Approximate Bayesian Computation in State Space ModelsPreprint
Martin, G. M., McCabe, B. P. M., Maneesoonthorn, W. O., & Robert, C. P. (2014) · arXiv (Cornell University) · doi:10.26180/21521130 · Cited by 20
Approximate Bayesian inference for complex ecosystems
Stumpf, M. P. H. (2014) · F1000Prime Reports · doi:10.12703/p6-60 · Cited by 19
“In parallel to their role in computationally demanding applications, ABC techniques have, more recently, also attracted attention as an inferential framework in their own right [16,51].”
BackgroundMethodSemi-automatic Approximate Bayesian ComputationPreprint
Fearnhead, P., & Prangle, D. (2010) · arXiv (Cornell University) · Cited by 19
Modelling Extremes Using Approximate Bayesian Computation
Erhardt, R. J., & Sisson, S. A. (2016) · Chapman & Hall/CRC eBooks · doi:10.1201/b19721-18 · Cited by 18
Köpke, C., Irving, J., & Roubinet, D. (2018) · Journal of Hydrology · doi:10.1016/j.jhydrol.2018.12.016 · Cited by 18
“…in recent years (e.g., Beven and Freer , 2001; Beven and Binley ,279 2014; Blasone et al., 2008; Nott et al., 2012; Sadegh and Vrugt , 2013a,b; Wilkinson, 2013),280 as researchers have realized the shortcomings of placing too much importance on the detailed281 statistical properties of the…”
BackgroundPeters, G. W., Kannan, B., Lasscock, B., Mellen, C., & Godsill, S. (2011) · Bayesian Analysis · doi:10.1214/11-ba628 · Cited by 18
Vakilzadeh, M. K., Beck, J. L., & Abrahamsson, T. J. S. (2018) · SIAM Journal on Scientific Computing · doi:10.1137/16m1090466 · Cited by 18
Informed sub-sampling MCMC: approximate Bayesian inference for large datasets
Maire, F., Friel, N., & Alquier, P. (2018) · Statistics and Computing · doi:10.1007/s11222-018-9817-3 · Cited by 18
“Overcoming situations where { Y = Z } is impossible or very unlikely has already been addressed in the ABC literature [see Fearnhead and Prangle (2012) and Wilkinson (2013)], leading to approximate ABC algorithms.”
BackgroundA First Course in Monte Carlo MethodsPreprint
Sanz-Alonso, D., & Ghattas, O. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2405.16359 · Cited by 18
Analyzing allele specific RNA expression using mixture models
Lü, R., Smith, R. M., Seweryn, M. T., Wang, D., Hartmann, K., Webb, A., Sadée, W., & Rempała, G. A. (2015) · BMC Genomics · doi:10.1186/s12864-015-1749-0 · Cited by 17
“For fitting the folded Skellam, we used a likelihood-free Markov chain Monte Carlo (MCMC) method [36], which can be also viewed as an Approximate Bayesian Computation (ABC) type of method [37].”
MethodConditional random matrix ensembles and the stability of dynamical systems
Kirk, P., Rolando, D., MacLean, A. L., & Stumpf, M. P. H. (2015) · New Journal of Physics · doi:10.1088/1367-2630/17/8/083025 · Cited by 18
Robust Bayesian Inference via Coarsening
Miller, J. W., & Dunson, D. B. (2018) · Figshare · doi:10.6084/m9.figshare.6220325 · Cited by 17
An overview on Approximate Bayesian computation
Baragatti, M., & Pudlo, P. (2014) · ESAIM Proceedings · doi:10.1051/proc/201444018 · Cited by 16
“As in Wilkinson (2013), one might also consider that the observed summary statistics are noisy, either because the (intractable) likelihood comes from a model that does not fit to the reality or because their data collection process has added some noise.”
BackgroundA note on target distribution ambiguity of likelihood-free samplersPreprint
Sisson, S. A., Peters, G. W., Briers, M., & Fan, Y. (2010) · arXiv (Cornell University) · doi:10.48550/arxiv.1005.5201 · Cited by 16
Hamiltonian ABCPreprint
Meeds, E., Leenders, R., & Welling, M. (2015) · arXiv (Cornell University) · doi:10.48550/arxiv.1503.01916 · Cited by 16
“In ABC Markov chain Monte Carlo (MCMC) [13, 26] the Metropolis-Hastings (MH) proposal distribution is composed of the product of the proposal for the parameters θ and the proposal for the simulator outputs:”
BackgroundHoffimann, J., Bufe, A., & Caers, J. (2019) · Journal of Geophysical Research Earth Surface · doi:10.1029/2019jf005245 · Cited by 15
“The approximation error associated with choosing a positive 𝜖 as opposed to zero is only understood in very simple situations where one assumes independent and identically distributed additive error (Van Der Vaart et al., 2018; Wilkinson, 2013).”
BackgroundTouloupou, P., Retkutė, R., Hollingsworth, T. D., & Spencer, S. E. F. (2020) · Spatial and Spatio-temporal Epidemiology · doi:10.1016/j.sste.2020.100391 · Cited by 15
Identifying cell-to-cell variability in internalization using flow cytometry
Browning, A. P., Ansari, N., Drovandi, C., Johnston, A. P. R., Simpson, M. J., & Jenner, A. L. (2022) · Journal of The Royal Society Interface · doi:10.1098/rsif.2022.0019 · Cited by 15
“Firstly, ABC is robust to model error, incorporating uncertainty due to factors that are not explicitly modelled [52] by approximating the likelihood through an acceptance criterion that allows for an imperfect (i.”
BackgroundAutomatic Variational ABCPreprint
Moreno, A., Adel, T., Meeds, E., Rehg, J. M., & Welling, M. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1606.08549 · Cited by 15
“Wilkinson [20] showed that in the likelihood-free setting, could be interpreted as model or simulation noise.”
BackgroundSimpson, M. J., & Plank, M. J. (2025) · Journal of The Royal Society Interface · doi:10.1098/rsif.2025.0536 · Cited by 14
Taking error into account when fitting models using Approximate Bayesian Computation
van der Vaart, E., Prangle, D., & Sibly, R. M. (2017) · Ecological Applications · doi:10.1002/eap.1656 · Cited by 14
“We do this using Wilkinson (2013)’s insight that accurate estimates can be obtained if the form of the error – the distribution of the differences between model outputs and data – is incorporated into the ABC protocol.”
InfluentialEstimation of Flow-Accelerated Corrosion Rate in Nuclear Piping System
Hazra, I., Pandey, M. D., & Jyrkama, M. I. (2019) · Journal of Nuclear Engineering and Radiation Science · doi:10.1115/1.4044407 · Cited by 14
The ensemble Kalman filter is an ABC algorithm
Nott, D. J., Marshall, L. A., & Ngoc, T. M. (2011) · Statistics and Computing · doi:10.1007/s11222-011-9300-x · Cited by 14
“Wilkinson (2008) shows how to interpret rejection type ABC algorithms as giving exact results when a model error term is taken into account.”
BackgroundApproximate Inference for Observation-Driven Time Series Models with Intractable Likelihoods
Jasra, A., Kantas, N., & Ehrlich, E. (2014) · ACM Transactions on Modeling and Computer Simulation · doi:10.1145/2592254 · Cited by 13
McKinley, T. J., Neal, P. J., Spencer, S. E. F., Conlan, A. J. K., & Tiley, L. S. (2019) · Bayesian Analysis · doi:10.1214/19-ba1174 · Cited by 13
“Recently Wilkinson (2013) provides an alternative interpretation of this approximate distribution as the exact posterior distribution for a model incorporating specific model discrepancy (or measurement error) terms.”
BackgroundA Likelihood-Free Reverse Sampler of the Posterior Distribution
Forneron, J.-J., & Ng, S. (2016) · Emerald Group Publishing Limited eBooks · doi:10.1108/s0731-905320160000036020 · Cited by 13
“b 1 () with one simulation (hence the notation Jb 1 ()), but solves it many times to obtain a distribution for b. The sequence of bare independent optimizers and do not have a Markov structure. 1 Wilkinson (2008) takes K ( ) to be the distribution of the error, or discrepancy, between the data and the best prediction…”
BackgroundMcGinty, R. J., Balick, D. J., Mirkin, S. M., & Sunyaev, S. R. (2025) · Nature Communications · doi:10.1038/s41467-025-66725-5 · Cited by 12
“Theapproximateposteriorisproportionaltothe product of the prior and the probability of acceptance under an ABC rejection strategy 55 .”
Andrieu, C., Lee, A., & Vihola, M. (2018) · Chapman & Hall/CRC eBooks · doi:10.1201/9781315117195-9 · Cited by 12
Lezcano-Casado, M., Baydin, A. G., Martínez-Rubio, D., Lê, T. A., Wood, F., Heinrich, L., Louppe, G., Cranmer, K. S., Ng, K., Bhimji, W., et al. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1712.07901 · Cited by 12
“Alternatively, approximate Bayesian computation (ABC) [24, 22] refers to a large class of approaches for sampling from the posterior distribution of these likelihood-free models where the original simulator is used directly as part of the inference engine.”
BackgroundWhy approximate Bayesian computational (ABC) methods cannot handle model choice problemsPreprint
Robert, C. P., Marin, J., & Pillai, N. S. (2011) · arXiv (Cornell University) · doi:10.48550/arxiv.1101.5091 · Cited by 12
“Users of ABC algorithms are therefore prepared for some degree of imprecision in their final answer, a point stressed by Wilkinson (2008) or Fearnhead and Prangle (2010) when they qualify ABC as exact inference on a wrong model.”
BackgroundMethodPreston, S. P., Wilkinson, R. D., Clayton, R. H., Chappell, M. J., & Mirams, G. R. (2025) · Current Opinion in Systems Biology · doi:10.1016/j.coisb.2025.100563 · Cited by 11
“Methods such as history matching (Vernon et al., 2014; Craig et al., 1997) and approximate Bayesian computation (Csill´ery et al., 2010; Wilkinson, 2013) are also useful tools for checking practical identifiability, and can incorporate an assessment of model discrepancy without requiring a full…”
Self-citationReply to Robert et al.: Model criticism informs model choice and model comparison
Ratmann, O., Andrieu, C., Wiuf, C., & Richardson, S. T. (2010) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.0912887107 · Cited by 11
Chen, Y. (2020) · Journal of the American Statistical Association · doi:10.1080/01621459.2020.1846974 · Cited by 11
ABC for climate: dealing with expensive simulatorsPreprint
Holden, P. B., Edwards, N. R., Hensman, J. J., & Wilkinson, R. D. (2015) · arXiv (Cornell University) · doi:10.48550/arxiv.1511.03475 · Cited by 11
“roximatingthedistribution ˇ ABC(jC obs) / Z I(ˆ(C obs;C sim) )ˇ(C simj)dC simˇ(): (2) TheacceptancekernelI(ˆ(C sim;C obs) ) implicitlyimpliesauniformdistribution for the simulator discrepancy (Wilkinson, 2013), but this is usually viewedasapragmaticcompromise,ratherthanamodellingdecision. An alternative to a probabili…”
BackgroundInfluentialSelf-citationBayesian and Algebraic Strategies to Design in Synthetic Biology
Araujo, R. P., Vittadello, S. T., & Stumpf, M. P. H. (2021) · Proceedings of the IEEE · doi:10.1109/jproc.2021.3129527 · Cited by 10
“Of these, approximate Bayesian computation [43], [44], or ABC, is perhaps the best known alternative to evaluating the full posterior via the likelihood function.”
Seeman, J., Slavković, A., & Reimherr, M. L. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-57521-2_23 · Cited by 10
Augustin, D., Lambert, B. C., Wang, K., Walz, A., Robinson, M. J., & Gavaghan, D. J. (2023) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1011135 · Cited by 10
Fearnhead, P. (2018) · Chapman & Hall/CRC eBooks · doi:10.1201/9781315117195-10 · Cited by 10
“ional Gaussian random variable with covariance matrix A( 0). Furthermore, for close to 0 the prior will be well approximated by a uniform prior. For the following we assume that D 0 is of rank p. Wilkinson (2013) shows that the eect of the approximation in ABC, whereby we accept simulated summaries which are similar,…”
MethodUnderstanding hormonal crosstalk in Arabidopsis root development via emulation and history matching
Jackson, S. E., Vernon, I., Liu, J., & Lindsey, K. (2020) · Statistical Applications in Genetics and Molecular Biology · doi:10.1515/sagmb-2018-0053 · Cited by 10
“There are other alternatives to history matching or a full Bayesian analysis, including the popular approach of Approximate Bayesian Computation (ABC) (Smith & Gelfand, 1992; Wilkinson, 2013), which could also be useful in this context.”
MethodZuluaga, C. D., Valencia, E. A., Álvarez, M. A., & Orozco, A. A. (2015) · Lecture notes in computer science · doi:10.1007/978-3-319-23231-7_5 · Cited by 9
“According to [17], there are different open questions around the ABC algorithm including how to choose the measure function ρ; what should be the value of ?; how to select the summary statistics s? ρ and are experimental and implementation issues.”
BackgroundMethodFilipe, J. A. N., & Kyriazakis, I. (2019) · Frontiers in Genetics · doi:10.3389/fgene.2019.00727 · Cited by 10
“First, when used as an approximation to a problem with a known data distribution, ABC produces a sample from a PD that is not the exact one supported by the likelihood as the tolerance parameter is non-zero (Robert et al., 2011; Wilkinson, 2013; Kypraios et al., 2017).”
MethodAn ABC approach for CAViaR models with asymmetric kernels
Tsiotas, G. (2020) · Journal of Statistical Computation and Simulation · doi:10.1080/00949655.2020.1727477 · Cited by 9
“Also, Wilkinson [29] has adopted a Gaussian process (GP) approach to accelerate ABC simulations by emulating a likelihood function.”
MethodInfluentialEstimation of Stochastic Degradation Models Using Uncertain Inspection Data
Lu, D. (2012) · UWSpace (University of Waterloo) · Cited by 9
Transflow Learning: Repurposing Flow Models Without RetrainingPreprint
Gambardella, A., Baydin, A. G., & Torr, P. H. S. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1911.13270 · Cited by 9
Järvenpää, M., Vehtari, A., & Marttinen, P. (2019) · arXiv · doi:10.48550/arxiv.1910.06121 · Cited by 8
“Standard ABC algorithms such as the ABC rejection sampler instead target the approximate posterior where Other choices of kernel π ε are also possible (Wilkinson, 2013).”
BackgroundMethodInformation, Entropy and Their Geometric Structures
Barbaresco, F., & Mohammad-Djafari, A. (2015) · MDPI eBooks · doi:10.3390/books978-3-03842-104-7 · Cited by 8
Méthode et outils pour l'identification de défauts des bâtiments connectés performants
Josse, R. (2017) · theses.fr (ABES) · doi:10.70675/0694011ez955ez4682z9ec3z56418635cdb1 · Cited by 8
Likelihood-free Model ChoicePreprint
Marin, J., Pudlo, P., Estoup, A., & Robert, C. P. (2015) · arXiv (Cornell University) · doi:10.48550/arxiv.1503.07689 · Cited by 8
“ows to innity, but we also benet from new results that set actual ABC implementations, with their nite computing power and strictly positive tolerances, within the range of other types of inference [38,39,40]. ABC now stands as an inference method that is justiable on its own ground. This approach may be the only solu…”
BackgroundABC SamplersPreprint
Fan, Y., & Sisson, S. A. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1802.09650 · Cited by 8
“This is more in line with the ABC approximation π̇ABC(θ|sobs) given by (9), and the interpretation of the ABC approximation to π(θ|sobs) as an exact model in the presence of model error due to Wilkinson (2013).”
MethodApproximate Bayesian Inference Based on Expected Evaluation
Hammer, H. L., Riegler, M. A., & Tjelmeland, H. (2023) · Bayesian Analysis · doi:10.1214/23-ba1368 · Cited by 7
One Pseudo-Sample is Enough in Approximate Bayesian Computation MCMC
Bornn, L., Pillai, N. S., Smith, A. D., & Woodard, D. B. (2014) · Cited by 7
Approximate Bayesian Computation for Granular and Molecular Dynamics Simulations
Kulakova, L., Angelikopoulos, P., Hadjidoukas, P. E., Papadimitriou, C., & Koumoutsakos, P. (2016) · Platform for Advanced Scientific Computing Conference (PASC) · doi:10.1145/2929908.2929918 · Cited by 7
“[40] R. D. D. Wilkinson.”
BackgroundMethodInfluentialMahrouz, N. (2020) · Journal of Hydrologic Engineering · doi:10.1061/(asce)he.1943-5584.0002048 · Cited by 7
Simulation-based marginal likelihood for cluster strong lensing cosmology
Killedar, M., Borgani, S., Fabjan, D., Dolag, K., Granato, G. L., Meneghetti, M., Planelles, S., & Ragone-Figueroa, C. (2017) · Monthly Notices of the Royal Astronomical Society · doi:10.1093/mnras/stx2262 · Cited by 7
Distilling importance samplingPreprint
Prangle, D. (2019) · arXiv (Cornell University) · Cited by 7
Practical parameter identifiability for spatiotemporal models of cell invasionPreprint
Simpson, M. J., Baker, R. E., Vittadello, S. T., & Maclaren, O. J. (2019) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/839282 · Cited by 7
Embedded model form uncertainty quantification with measurement noise for Bayesian model calibration
Arcones, D. A., Weiser, M., Koutsourelakis, P., & Unger, J. F. (2026) · Data-Centric Engineering · doi:10.1017/dce.2025.10035 · Cited by 6
“Finally, the modification of the acceptance criteria has been proven to correctly converge to the posterior of the summary statistics with noise (Wilkinson (2013)).”
InfluentialApproximate Bayesian Computation by Subset Simulation for model selection in dynamical systems
Vakilzadeh, M. K., Beck, J. L., & Abrahamsson, T. J. S. (2017) · Procedia Engineering · doi:10.1016/j.proeng.2017.09.291 · Cited by 6
“Wilkinson [9] also showed that a standard ABC posterior gives an exact posterior distribution for a new model under the assumption that the summary statistics are corrupted with a uniform additive error term.”
BackgroundBayesian Verification of Chemical Reaction Networks
Molyneux, G. W., Wijesuriya, V. B., & Abate, A. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-54997-8_29 · Cited by 6
Investigating Uncertainty in Global Hydrology Modelling
Smith, K. (2015) · Nottingham ePrints (University of Nottingham) · doi:10.13140/rg.2.2.22810.95685 · Cited by 7
Data adaptation in HANDY economy-ideology modelPreprint
Sendera, M. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1904.04309 · Cited by 6
$γ$-ABC: Outlier-Robust Approximate Bayesian Computation Based on a Robust Divergence EstimatorPreprint
Fujisawa, M., Teshima, T., Sato, I., & Sugiyama, M. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2006.07571 · Cited by 6
“1 Gaussian Mixture Model (GM) The univariate gaussian mixture model is the most fundamental benchmark method in ABC literature [51, 58, 31].”
BackgroundMethodCalibrating high frequency trading data to agent based models using approximate Bayesian computation
Goosen, K. (2021) · Open University of Cape Town (University of Cape Town) · Cited by 6
Summary statistics and sequential methods for approximate Bayesian computation
Prangle, D. (2011) · Lancaster EPrints (Lancaster University) · Cited by 6
A Cautionary Tale of Model Misspecification and Identifiability
Browning, A. P., Flegg, J. A., & Murphy, R. J. (2025) · Bulletin of Mathematical Biology · doi:10.1007/s11538-025-01573-4 · Cited by 5
“Approximate Bayesian computation methods [53, 54] (in addition to those based on a neural likelihood [55], and control theory [56]) are also able to account for model misspecification, by relaxing the typical requirement that the model must fit in the infinite data limit.”
On predictive inference for intractable models via approximate Bayesian computation
Järvenpää, M., & Corander, J. (2023) · Statistics and Computing · doi:10.1007/s11222-022-10163-6 · Cited by 5
“Wilkinson (2013) has shown that these approaches can be interpreted as different measurement error models.”
BackgroundTime Series of Counts under Censoring: A Bayesian Approach
Silva, I. V. D. A., Silva, M. E., Pereira, I., & McCabe, B. P. M. (2023) · Entropy · doi:10.3390/e25040549 · Cited by 5
“The most basic approximate acceptance/rejection algorithm, based on the works of [17,18], is as follows: 1.”
MethodApproximate Probabilistic Power Flow
Zuluaga, C. D., & Álvarez, M. A. (2017) · Lecture notes in computer science · doi:10.1007/978-3-319-50947-1_5 · Cited by 5
Applications of Probabilistic Programming (Master's thesis, 2015)Preprint
Perov, Y. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1606.00075 · Cited by 5
“(2.4) The important point, which was made by Wilkinson himself in (Wilkinson, 2013), is that if the model already includes the error as part of its generative model (i.e.”
BackgroundInfluentialApproximate Bayesian Computation via ClassificationPreprint
Wang, Y., Kaji, T., & Ročková, V. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2111.11507 · Cited by 5
“Wilkinson (2013) and Sisson et al. (2018) provide a thorough overview and comparisons of the commonly used kernels.”
BackgroundA Primer on Bayesian Parameter Estimation and Model Selection for Battery Simulators
Kuhn, Y., Adachi, M., Philipp, M., Howey, D. A., & Horstmann, B. (2026) · Journal of The Electrochemical Society · doi:10.1149/1945-7111/ae73f3 · Cited by 4
Molecular Reconstruction of Heavy Petroleum Fractions: A Bayesian Approach
Maciel, H. S., Fernandes, D. T., Abreu, C. R. A., Secchi, A. R., & Tavares, F. W. (2024) · Energy & Fuels · doi:10.1021/acs.energyfuels.4c00889 · Cited by 4
Moment-based parameter inference with error guarantees for stochastic reaction networks
Li, Z., Barahona, M., & Thomas, P. (2025) · The Journal of Chemical Physics · doi:10.1063/5.0251744 · Cited by 4
“For stochastic reaction networks, however, the computation of likelihoods relies on explicit solutions of the CME, a) Electronic mail: zekai.li18@imperial.ac.uk b) Corresponding author: p.thomas@imperial.ac.uk which quickly become intractable even for relatively simple networks 28,29 .”
MethodHybrid inspection-monitoring approach for optimal maintenance planning
Geara, C., Faddoul, R., Chateauneuf, A., & Raphael, W. E. (2020) · Structure and Infrastructure Engineering · doi:10.1080/15732479.2020.1713168 · Cited by 4
Dasgupta, A., & Johnson, E. A. (2024) · Journal of Engineering Mechanics · doi:10.1061/jenmdt.emeng-7204 · Cited by 4
Wilkinson, D. J. (2011) · Wiley eBooks · doi:10.1002/9781119970606.ch18 · Cited by 4
Metadata
- Authors
- Wilkinson, R. D.
- Affiliations
- University of Nottingham (GB)
- Corresponding author
- Wilkinson, Richard David
- Venue
- Statistical Applications in Genetics and Molecular Biology
- Publisher
- Walter de Gruyter GmbH
- Published
- 2013-01-06
- Volume
- 12
- Issue
- 2
- Pages
- 129-41
- ISSN
- 1544-6115 (ISSN-L)
- 2194-6302
- OpenAlex
- W2053332694
- Cited by
- 247
- References
- 41
- Citations by year
- 2026: 6 · 2025: 9 · 2024: 8 · 2023: 8 · 2022: 9 · 2021: 13 · 2020: 30 · 2019: 16 · 2018: 26 · 2017: 13 · 2016: 17 · 2015: 24 · 2014: 17 · 2013: 16
- FWCI
- 21.70
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- Statistical Applications in Genetics and Molecular Biology (publishedVersion)
- PubMed (publishedVersion)
- White Rose Research Online (University of Leeds, The University of Sheffield, University of York) (acceptedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Topics
- Physical Sciences › Mathematics › Statistics and Probability › Markov Chains and Monte Carlo Methods
- Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods and Bayesian Inference
- Physical Sciences › Computer Science › Artificial Intelligence › Gaussian Processes and Bayesian Inference
- Keywords
- approximate Bayesian computation, model error, simulation-based inference, acceptance probability, model calibration
- MeSH terms
- Algorithms, Bayes Theorem, Computer Simulation, Markov Chains, Monte Carlo Method, Models, Statistical, Likelihood Functions
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.