A Simulated Annealing Approach to Approximate Bayes Computations

TL;DR

This paper introduces a novel particle-based algorithm for Approximate Bayes Computations inspired by Simulated Annealing, which avoids importance sampling and adaptively optimizes convergence to the posterior.

Contribution

It presents a new class of particle algorithms for ABC that do not rely on importance sampling and includes an adaptive scheme for tolerance and jump distribution optimization.

Findings

Algorithm converges under specified tolerance decrease conditions.

Adaptive scheme improves convergence efficiency.

Compared favorably against existing algorithms on toy examples.

Abstract

Approximate Bayes Computations (ABC) are used for parameter inference when the likelihood function of the model is expensive to evaluate but relatively cheap to sample from. In particle ABC, an ensemble of particles in the product space of model outputs and parameters is propagated in such a way that its output marginal approaches a delta function at the data and its parameter marginal approaches the posterior distribution. Inspired by Simulated Annealing, we present a new class of particle algorithms for ABC, based on a sequence of Metropolis kernels, associated with a decreasing sequence of tolerances w.r.t. the data. Unlike other algorithms, our class of algorithms is not based on importance sampling. Hence, it does not suffer from a loss of effective sample size due to re-sampling. We prove convergence under a condition on the speed at which the tolerance is decreased. Furthermore, we present a scheme that adapts the tolerance and the jump distribution in parameter space according to some mean-fields of the ensemble, which preserves the statistical independence of the particles, in the limit of infinite sample size. This adaptive scheme aims at converging as close as possible to the correct result with as few system updates as possible via minimizing the entropy production in the system. The performance of this new class of algorithms is compared against two other recent algorithms on two toy examples.