Expectation Consistent Approximate Inference

By Manfred Opper, Ole Winther. 2005

248 citations recorded by the source corpus.

How researchers see this paper

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

Citations in the corpus

Variational Inference: A Review for Statisticians (2016, background): “While alternative divergences may be difficult to optimize, they may give better approximations (Minka 2005; Opper and Winther 2005).”
Black-box $\alpha$-divergence Minimization (2015, methodology): “On the second issue, Heskes & Zoeter (2002) and Opper & Winther (2005) derived a convergent double-loop implementation of EP.”
Expectation propagation for neural networks with sparsity-promoting priors (2013, background): “The EP approximation logZEP has the appealing property that its partial derivatives with respect to the site parameters in their canonical forms5 are zero when the algorithm has been iterated until convergence (Opper and Winther, 2005).”
Gaussian Kullback-Leibler approximate inference (2013, background): “Provably convergent double loop extensions to G-EP have been developed—see Opper and Winther (2005) and references therein for details.”
Rényi Divergence Variational Inference (2016, background): “This paper focuses on optimisation-based approximate inference algorithms, popular examples of which include variational inference (VI), variational Bayes (VB) [1, 2] and expectation propagation (EP) [3, 4].”
Graphical Models, Exponential Families, and Variational Inference (2008, methodology): “As with assumed density filtering, expectationpropagation [172, 171] and various related algorithms [181, 61, 112] are typically described in terms of moment-matching operations.”
Approximations for Binary Gaussian Process Classification (2008, methodology): “The lower bound provided by Jensen's inequality Z B (Equation 9) is known to be below the approximation Z EP obtained by EP (Opper and Winther, 2005, page 2183).”
Approximate Gaussian Integration using Expectation Propagation (2011, background): “Second, addressing convergence issues via a provably convergent double-loop approach to EP (Opper and Winther, 2005; Seeger and Nickish, 2011).”
Stochastic Expectation Propagation (2015, methodology): “One family of approximation method has garnered less attention in this regard: Expectation Propagation (EP) [7, 8].”
Variational Inference with Tail-adaptive f-Divergence (2018, methodology): “KL( q || p ) is the typical objective function used in variational inference; the reversed direction KL( p || q ) is also used in various settings (e.g., Minka, 2001; Opper & Winther, 2005; De Boer et al., 2005; Burda et al., 2016).”
Vector Approximate Message Passing (2019, methodology): “Among these are the Adaptive Thouless-Anderson-Palmer (ADATAP) [12], Expectation Propagation (EP) [13], [14], Expectation Consistent Approximation (EC) [15]–[17], (S-transform AMP) S-AMP [18], [19], and (Orthogonal AMP) OAMP [20] approaches.”
Expectation Propagation as a Way of Life: A Framework for Bayesian Inference on Partitioned Data (2014, methodology): “In its general form, the problem can be formulated as an optimization for the free energy corresponding to the negative logarithm of the intractable normalizer of the global approximation in Equation (1) (Opper and Winther, 2005).”