Approximate Bayesian Computation

By MARK A. Beaumont. Annual Review of Statistics and Its Application, 2019

595 citations recorded by the source corpus.

Many of the statistical models that could provide an accurate, interesting, and testable explanation for the structure of a data set turn out to have intractable likelihood functions. The method of approximate Bayesian computation (ABC) has become a popular approach for tackling such models. This review gives an overview of the method and the main issues and challenges that are the subject of current research.

Subfields: Artificial Intelligence, Mathematical Physics, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/e7e144b62dde504f8996a459e261261a984ef38e

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Mathematical Physics

In Statistics and Probability

Citations in the corpus

fastsimcoal2: demographic inference under complex evolutionary scenarios (2021, methodology): “…to their versatility, they have also been used in Approximate Bayesian Computations (ABC, Beaumont et al. , 2002) to estimate parameters under very complex models and to perform model testing (Beaumont, 2019; Currat et al. , 2019; Mondal et al. , 2019; Sanchez et al. , 2020; Wegmann et al. , 2010).”
Asymptotic properties of approximate Bayesian computation (2016, methodology): “g=1 [ p̂gαj{μ|η(y)} − p (μ|y) ]2)1/2 , (15)”
BayesFlow: Learning Complex Stochastic Models With Invertible Neural Networks (2020, methodology): “suitable statistics for each application a priori [36, 35, 46, 49].”
Inferring brain-computational mechanisms with models of activity measurements (2016, methodology): “This multinormal model enables us to evaluate the likelihood (probability of a crossnobis RDM estimate given a BCM and MM) and thus obviates the need to model the effect of noise by simulation or to rely on an approximate Bayesian computation (Sunnåker et al. 2013; Robert et al. 2011).”
Efficient Probabilistic Inference in the Quest for Physics Beyond the Standard Model (2018, unknown): “Alternatively, approximate Bayesian computation (ABC) [12, 13] 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.”
Inverse Reward Design (2017, methodology): “This approach, inspired by methods in approximate Bayesian computation (Sunnåker et al., 2013), samples a ﬁnite set of weights { w i } to approximate the integral in Equation 2.”
Analyzing Inverse Problems with Invertible Neural Networks (2018, methodology): “If a model y = s(x) for the forward process is available, approximate Bayesian computation is often preferred, which embeds the forward model in a rejection sampling scheme for the posterior p(x|y) [40, 26, 46].”