Nonparametric Bayesian Density Modeling with Gaussian Processes

We present the Gaussian process density sampler (GPDS), an exchangeable generative model for use in nonparametric Bayesian density estimation. Samples drawn from the GPDS are consistent with exact, independent samples from a distribution defined by a density that is a transformation of a function drawn from a Gaussian process prior. Our formulation allows us to infer an unknown density from data using Markov chain Monte Carlo, which gives samples from the posterior distribution over density functions and from the predictive distribution on data space. We describe two such MCMC methods. Both methods also allow inference of the hyperparameters of the Gaussian process.

Image deconvolution using a Gaussian scale mixtures model to approximate the wavelet sparseness constraint

This paper proposes to use an extended Gaussian Scale Mixtures (GSM) model instead of the conventional ℓ1 norm to approximate the sparseness constraint in the wavelet domain. We combine this new constraint with subband-dependent minimization to formulate an iterative algorithm on two shift-invariant wavelet transforms, the Shannon wavelet transform and dual-tree complex wavelet transform (DTCWT). This extented GSM model introduces spatially varying information into the deconvolution process and thus enables the algorithm to achieve better results with fewer iterations in our experiments. ©2009 IEEE.

Creep response of sandwich beams with a metallic foam core

The creep response of metallic foam sandwich beams in 3-point bend is investigated numerically for the case of a metallic foam core and two steel faces. The face sheets are treated as elastic, while the foam core is modeled by a viscoplastic extension of the Deshpande-Fleck yield surface. This power-law creeping constitutive law has been implemented within the commercial finite element code ABAQUS. It is found that the beams creep by a variety of competing mechanisms, depending upon the choice of material properties and the geometric parameters. A failure map is constructed and effect of rate dependence on the load-deflection curves is quantified, and compared against the available experimental data.

Plastic collapse of sandwich beams with a metallic foam core

Plastic collapse modes of sandwich beams have been investigated experimentally and theoretically for the case of an aluminum alloy foam with cold-worked aluminum face sheets. Plastic collapse is by three competing mechanisms: face yield, indentation and core shear, with the active mechanism depending upon the choice of geometry and material properties. The collapse loads, as predicted by simple upper bound solutions for a rigid, ideally plastic beam, and by more refined finite element calculations are generally in good agreement with the measured strengths. However, a thickness effect of the foam core on the collapse strength is observed for collapse by core shear: the shear strength of the core increases with diminishing core thickness in relation to the cell size. Limit load solutions are used to construct collapse maps, with the beam geometrical parameters as axes. Upon displaying the collapse load for each collapse mechanism, the regimes of dominance of each mechanism and the associate mass of the beam are determined. The map is then used in optimal design by minimizing the beam weight for a given structural load index.

Quasi-static three-point bending of carbon fiber sandwich beams with square honeycomb cores

Sandwich beams comprising identical face sheets and a square honeycomb core were manufactured from carbon fiber composite sheets. Analytical expressions were derived for four competing collapse mechanisms of simply supported and clamped sandwich beams in three-point bending: core shear, face microbuckling, face wrinkling, and indentation. Selected geometries of sandwich beams were tested to illustrate these collapse modes, with good agreement between analytic predictions and measurements of the failure load. Finite element (FE) simulations of the three-point bending responses of these beams were also conducted by constructing a FE model by laying up unidirectional plies in appropriate orientations. The initiation and growth of damage in the laminates were included in the FE calculations. With this embellishment, the FE model was able to predict the measured load versus displacement response and the failure sequence in each of the composite beams. © 2011 American Society of Mechanical Engineers.

Efficient line launch for bandwidth improvement of 10 Gbit/s multimode fibre links using elliptical Gaussian beam

A novel elliptical Gaussian beam line launch is shown to allow improved system capacity for high speed multimode fibre links. This launch maintains higher bandwidth than dual launch even with misalignment of ±6 μm despite not requiring testing at installation. ©2010 IEEE.

Tractable nonparametric bayesian inference in poisson processes with Gaussian process intensities

The inhomogeneous Poisson process is a point process that has varying intensity across its domain (usually time or space). For nonparametric Bayesian modeling, the Gaussian process is a useful way to place a prior distribution on this intensity. The combination of a Poisson process and GP is known as a Gaussian Cox process, or doubly-stochastic Poisson process. Likelihood-based inference in these models requires an intractable integral over an infinite-dimensional random function. In this paper we present the first approach to Gaussian Cox processes in which it is possible to perform inference without introducing approximations or finitedimensional proxy distributions. We call our method the Sigmoidal Gaussian Cox Process, which uses a generative model for Poisson data to enable tractable inference via Markov chain Monte Carlo. We compare our methods to competing methods on synthetic data and apply it to several real-world data sets. Copyright 2009.

Tractable nonparametric Bayesian inference in Poisson processes with Gaussian process intensities

The inhomogeneous Poisson process is a point process that has varying intensity across its domain (usually time or space). For nonparametric Bayesian modeling, the Gaussian process is a useful way to place a prior distribution on this intensity. The combination of a Poisson process and GP is known as a Gaussian Cox process, or doubly-stochastic Poisson process. Likelihood-based inference in these models requires an intractable integral over an infinite-dimensional random function. In this paper we present the first approach to Gaussian Cox processes in which it is possible to perform inference without introducing approximations or finite-dimensional proxy distributions. We call our method the Sigmoidal Gaussian Cox Process, which uses a generative model for Poisson data to enable tractable inference via Markov chain Monte Carlo. We compare our methods to competing methods on synthetic data and apply it to several real-world data sets.

Symmetrical 2-D Hermite-Gaussian square launch for high bit rate transmission in multimode fiber links

A 2-D Hermite-Gaussian square launch is demonstrated to show improved systems capacity over multimode fiber links. It shows a bandwidth improvement over both center and offset launches and exhibits ±5 μm misalignment tolerance. © 2011 Optical Society of America.