Gaussian processes for machine learning (GPML) toolbox

The GPML toolbox provides a wide range of functionality for Gaussian process (GP) inference and prediction. GPs are specified by mean and covariance functions; we offer a library of simple mean and covariance functions and mechanisms to compose more complex ones. Several likelihood functions are supported including Gaussian and heavy-tailed for regression as well as others suitable for classification. Finally, a range of inference methods is provided, including exact and variational inference, Expectation Propagation, and Laplace’s method dealing with non-Gaussian likelihoods and FITC for dealing with large regression tasks.

Approximations for Binary Gaussian Process Classification

We provide a comprehensive overview of many recent algorithms for approximate inference in Gaussian process models for probabilistic binary classification. The relationships between several approaches are elucidated theoretically, and the properties of the different algorithms are corroborated by experimental results. We examine both 1) the quality of the predictive distributions and 2) the suitability of the different marginal likelihood approximations for model selection (selecting hyperparameters) and compare to a gold standard based on MCMC. Interestingly, some methods produce good predictive distributions although their marginal likelihood approximations are poor. Strong conclusions are drawn about the methods: The Expectation Propagation algorithm is almost always the method of choice unless the computational budget is very tight. We also extend existing methods in various ways, and provide unifying code implementing all approaches.

Design and fabrication of color filters for projection display system

Color filters are key components in an optical engine projection display system. In this paper, a new admittance-matching method for designing and fabricating the high performance filters is described, in which the optimized layers are limited to the interfaces between the stack (each combination of quarter-wave-optical-thickness film layers is called a stack) and stack, or between stack and substrate, or between stack and incident medium. This method works well in designing filters containing multiple stacks such as UV-IR cut and broadband filters. The tolerance and angle sensitivity for the designed film stacks are analyzed. The thermal stability of the sample color filters was measured. A good result in optical performance and thermal stability was obtained through the new design approach. (c) 2006 Society of Photo-Optical Instrumentation Engineers.

Electric field enhancement in guided-mode resonance filters

Electric fields inside guided-mode resonance filters (GMRFs) may be intensified by resonance effects. The electric field enhancement is investigated in two GMRFs: one is resonant at normal incidence, the other at oblique incidence. It is shown that the two GMRFs exhibit different behaviors in their electric enhancement. Differences between the electric field distributions of the two GMRFs arise because coupling between counter-propagating modes occurs in the first case. It is also shown that the order of the electric field of maximum amplitude can be controlled by modulation of the dielectric constant of the grating. (c) 2006 Optical Society of America.

Far-field intensity distribution and M-2 factor of hollow Gaussian beams

The far-field intensity distribution of hollow Gaussian beams was investigated based on scalar diffraction theory. An analytical expression of the M-2 factor of the beams was derived on the basis of the second-order moments. Moreover, numerical examples to illustrate our analytical results are given. (c) 2005 Optical Society of America.

Far-field intensity distribution of beam generated by Gaussian mirror resonator

Based on scalar diffraction theory, we investigated far-field intensity distribution (FFID) of beam generated by Gaussian mirror resonator. We found usable analytical expressions of diffracted field with respect to variation of diffraction parameters. Particular attention was paid to the parameters such as mirror spot size and radius of the Gaussian mirror, which determine the FFID. All analyses were limited to TEM00 fundamental mode. (c) 2004 Elsevier B.V. All rights reserved.

Propagation properties of a hard-edged diffracted beam generated by a Gaussian mirror resonator

Based on the scalar diffraction theory, the propagation and focusing properties of a hard-edged diffracted beam generated by a Gaussian mirror resonator were investigated. Explicit expressions for the field distribution of the truncated beam that propagates through a paraxial optical ABCD system were derived in detail. Numerical examples are given to illustrate our analytical results. (c) 2006 Optical Society of America.

Propagation properties of beam generated by Gaussian mirror resonator

Based on scalar diffraction theory, propagation properties of beam generated by Gaussian mirror resonator were investigated. Explicit expressions for the field distribution of the beam that propagate through a paraxial optical ABCD system were derived and numerical examples were given to illustrate our analytical results. (c) 2005 Elsevier B.V. All rights reserved.