Variational inference (VI) combined with data subsampling enables approximate posterior inference with large data sets for otherwise intractable models, but suffers from poor local optima. We first formulate a deterministic annealing approach for the generic class of conditionally conjugate exponential family models. This algorithm uses a temperature parameter that deterministically deforms the...

We present an optimization approach to the weak approximation of a general class of stochastic differential equations with jumps, in particular, when value functions with compact support are considered. Our approach employs a mathematical programming technique yielding upper and lower bounds of the expectation, without Monte Carlo sample paths simulations, based upon the exponential tempering o...

D. A. Skobir Balantič, B. Žužek, F. Vodopivec, Institute of Metals and Technology, Ljubljana, Slovenia Specimens of high chromium creep resistant steel were tempered at 800 °C for different times and examined in SEM. After short tempering stringers of cementite particles are formed at ferrite grain boundaries. By longer tempering, the content of chromium and molybdenum increase up to Cr18Fe3Mo2...

We obtain upper bounds on the convergence rates of Markov chains constructed by parallel and simulated tempering. These bounds are used to provide a set of sufficient conditions for torpid mixing of both techniques. We apply these conditions to show torpid mixing of parallel and simulated tempering for three examples: a normal mixture model with unequal covariances in RM and the mean-field Pott...

beside developing strategies, success of organizations depends on alignment of strategies. one of the important functional areas in which companies develop strategies is operation. alignment of operation strategies is called horizontal alignment. the aim of this article is to present a model for alignment of operation strategies. to this end, an original algorithm has been presented by which ho...

The relative performances of different implementations of the Wang-Landau method are assessed on two classes of systems with continuous degrees of freedom, namely, two polypeptides and two atomic Lennard-Jones clusters. Parallel tempering Monte Carlo simulations serve as a reference, and we pay particular attention to the variations of the multiplicative factor f during the course of the simula...

We show that the acceptance probability for swaps in the parallel tempering Monte Carlo method for classical canonical systems is given by a universal function that depends on the average statistical fluctuations of the potential and on the ratio of the temperatures. The law, called the incomplete beta function law, is valid in the limit that the two temperatures involved in swaps are close to ...

For every v ≡ 5 (mod 6) there exists a pairwise balanced design (PBD) of order v with exactly one block of size 5 and the rest of size 3. We will refer to such a PBD as a PBD(5∗, 3). A flower in a PBD(5∗, 3) is the set of all blocks containing a given point. If (S,B) is a PBD(5∗, 3) and F is a flower, we will write F ∗ to indicate that F contains the block of size 5. The flower intersection pro...

Simulated tempering (ST) is an established Markov chain Monte Carlo (MCMC) method for sampling from a multimodal density π(θ). Typically, ST involves introducing an auxiliary variable k taking values in a finite subset of [0, 1] and indexing a set of tempered distributions, say πk(θ) ∝ π(θ) k. In this case, small values of k encourage better mixing, but samples from π are only obtained when the...

This article introduces a new variant on the popular optimization method, simulated annealing, herein called modified discrete-state simulated annealing. A guide in practical implementation of simulated tempering is also provided. The algorithms are tested in the design of an ideal collector for a travelling wave tube amplifier. keywords: simulated annealing, simulated tempering, travelling wav...