Gibbs sampling is a widely applicable inference technique that can in principle deal with complex multimodal distributions. Unfortunately, it fails in many practical applications due to slow convergence and abundance of local minima. In this paper, we propose a general method of speeding up Gibbs sampling in probabilistic models. The method works by introducing auxiliary variables which represe...

We consider the problem of inference in a probabilistic model for transient populations where we wish to learn about arrivals, departures, and population size over all time, but the only available data are periodic counts of the population size at specific observation times. The underlying model arises in queueing theory (as an Mt/G/∞ queue) and also in ecological models for short-lived animals...

Many scalable video coding systems use variable resolution frames to enable different decoding layers. Some of these systems also use frame down-sampling along with enhancement layers to reduce complexity. In order to do that, super-resolution methods associated with efficient interpolation processes may help to increase the quality of low-resolution frames. This work presents a super-resolutio...

We tackle the facial landmark localization problem as an inference problem over a Markov Random Field. Efficient inference is implemented using Gibbs sampling with approximated full conditional distributions in a latent variable model. This approximation allows us to improve the runtime performance 1000-fold over classical formulations with no perceptible loss in accuracy. The exceptional robus...

This work presents simple and fast structured Bayesian learning for matrix and tensor factorization models. An unblocked Gibbs sampler is proposed for factorization machines (FM) which are a general class of latent variable models subsuming matrix, tensor and many other factorization models. We empirically show on the large Netflix challenge dataset that Bayesian FM are fast, scalable and more ...

We describe a sampling technique particularly suitable for active vision: Dimensionally-Independent Exponential Mapping (DIEM), in which each dimension of the original data is sampled in an exponentially increasing or decreasing series of steps, with bilateral symmetry about the data mid-point. Multidimensional data sampling is achieved by combining single dimension sampling coordinates. DIEM i...