Bayesian Personalized Ranking (BPR) is a representative pairwise learningmethod for optimizing recommendationmodels. It is widely known that the performance of BPR depends largely on the quality of the negative sampler. In this short paper, we make two contributions with respect to BPR. First, we find that sampling negative items from the whole space is unnecessary and may even degrade the perf...

As linguistic models incorporate more subtle nuances of language and its structure, standard inference techniques can fall behind. These models are often tightly coupled such that they defy clever dynamic programming tricks. Here we demonstrate that Sequential Monte Carlo approaches, i.e. particle filters, are well suited to approximating such models. We implement two particle filters, which jo...

The Gibbs Centroid Sampler is a software package designed for locating conserved elements in biopolymer sequences. The Gibbs Centroid Sampler reports a centroid alignment, i.e. an alignment that has the minimum total distance to the set of samples chosen from the a posteriori probability distribution of transcription factor binding-site alignments. In so doing, it garners information from the f...

SUMMARY In this article we investigate the relationship between the two popular algorithms, the EM algorithm and the Gibbs sampler. We show that the approximate rate of convergence of the Gibbs sampler by Gaussian approximation is equal to that of the corresponding EM type algorithm. This helps in implementing either of the algorithms as improvement strategies for one algorithm can be directly ...

A fundamental task in machine learning and related fields is to perform inference on Bayesian networks. Since exact inference takes exponential time in general, a variety of approximate methods are used. Gibbs sampling is one of the most accurate approaches and provides unbiased samples from the posterior but it has historically been too expensive for large models. In this paper, we present an ...

The symmetric-range accuracy A of a sampler is defined as the fractional range, symmetric about the true concentration, that includes a specified proportion of sampler measurements. In this article, we give an explicit expression for A assuming that the sampler measurements follow a normal distribution. We propose confidence limits for A based on the concept of a 'generalized confidence interva...

A new framework for color image segmentation is introduced generalizing the concepts of point-based and spatially-based methods. This framework is based on Markov Random Fields using a Continuous Gibbs Sampler. The Markov Random Fields approach allows for a rigorous computational framework where local and global spatial constraints can be globally optimized. Using a Continuous Gibbs Sampler ena...

We present an MCMC sampler for Dirichlet process mixture models that can be parallelized to achieve significant computational gains. We combine a nonergodic, restricted Gibbs iteration with split/merge proposals in a manner that produces an ergodic Markov chain. Each cluster is augmented with two subclusters to construct likely split moves. Unlike some previous parallel samplers, the proposed s...

This paper is devoted to the construction of Boltzmann samplers according to various distributions, and uses stochastic bias on the parameter of a Boltzmann sampler, to produce a sampler with a different distribution for the size of the output. As a significant application, we produce Boltzmann samplers for words defined by regular specifications containing shuffle operators and linear recursio...

Abstract. An impulse sampler multiplies an input signal u by a periodic delta impulse train of period z. If z is small, then the output signal of the sampler (or filtered versions thereof) can be used as an approximation of u. If u belongs to the Sobolev space H s with s > 89 then the output is in H -s. Our main result is that as the sampling period z becomes small, the impulse sampler approxim...