Bayesian inference and learning in Gaussian process state-space models with Particle MCMC

State-space models are successfully used in many areas of science, engineering and economics to model time series and dynamical systems. We present a fully Bayesian approach to inference and learning (i.e. state estimation and system identification) in nonlinear nonparametric state-space models. We place a Gaussian process prior over the state transition dynamics, resulting in a flexible model able to capture complex dynamical phenomena. To enable efficient inference, we marginalize over the transition dynamics function and, instead, infer directly the joint smoothing distribution using specially tailored Particle Markov Chain Monte Carlo samplers. Once a sample from the smoothing distribution is computed, the state transition predictive distribution can be formulated analytically. Our approach preserves the full nonparametric expressivity of the model and can make use of sparse Gaussian processes to greatly reduce computational complexity.

Classification using log Gaussian Cox processes

McCullagh and Yang (2006) suggest a family of classification algorithms based on Cox processes. We further investigate the log Gaussian variant which has a number of appealing properties. Conditioned on the covariates, the distribution over labels is given by a type of conditional Markov random field. In the supervised case, computation of the predictive probability of a single test point scales linearly with the number of training points and the multiclass generalization is straightforward. We show new links between the supervised method and classical nonparametric methods. We give a detailed analysis of the pairwise graph representable Markov random field, which we use to extend the model to semi-supervised learning problems, and propose an inference method based on graph min-cuts. We give the first experimental analysis on supervised and semi-supervised datasets and show good empirical performance.

Z-scan theory based on a diffraction model

Based on Fresnel-Mrchhoff diffraction theory, a diffraction model of nonlinear optical media interacting with a Gaussian beam has been set up that can interpret the Z-scan phenomenon in a new way. This theory not only is consistent with the conventional Z-scan theory for a small nonlinear phase shift but also can be used for larger nonlinear phase shifts. Numerical computations indicate that the shape of the Z-scan curve is greatly affected by the value of the nonlinear phase shift. The symmetric dispersionlike Z-scan curve is valid only for small nonlinear p base shifts (\Deltaphi(0)\ < pi), but, with increasingly larger nonlinear phase shifts, the valley of the transmittance is severely suppressed and the peak is greatly enhanced. The power output through the aperture will oscillate with increasing nonlinear phase shift caused by the input laser power. The aperture transmittance will attenuate and saturate with increasing Kerr constant. (C) 2003 Optical Society of America.

Two-component model for the interstellar scattering in direction to the Vela pulsar

This paper discusses a rigorous treatment of the refractive scintillation of pulsar PSR B0833-45 caused by a two-component interstellar scattering medium. It is assumed that the interstellar scattering medium is composed of a thin screen ISM and an extended interstellar medium. We consider that the scattering of the thin screen concentrates in a thin layer presented by a delta function distribution and that the scattering density of the extended irregular medium satisfies the Gaussian distribution. We investigate and develop equations for the flux density structure function corresponding to this two-component ISM geometry in the scattering density distribution and compare our result with that of the Vela pulsar observations. We conclude that the refractive scintillation caused by this two-component ISM scattering gives a more satisfactory explanation for the observed flux density variation of the Vela pulsar than does the single extended medium model. The level of refractive scintillation is strongly sensitive to the distribution of scattering material along the line of sight. The logarithmic slope of the structure function is sensitive to thin screen location and is relatively insensitive to the scattering strength of the thin screen medium. Therefore, the proposed model can be applied to interpret the structure function of flux density observed in pulsar PSR B0833-45. The result suggests that the medium consists of a discontinuous distribution of plasma turbulence embedded in the Vela supernova remnant. Thus our work provides some insight into the distribution of the scattering along the line of sight to the Vela pulsar.

Retrieval of Electron Return Time from High-order Harmonics Generated in a Mixture of He and Ne Gases

In terms of single-atom induced dipole moment by Lewenstein model, we present the macroscopic high-order harmonic generation from mixed He and Ne gases with different mixture ratios by solving three-dimensional Maxwell's equation of harmonic field. And then we show the validity of mixture formulation by Wagner et al. [Phys. Rev. A 76 (2007) 061403(R)] in macroscopic response level. Finally, using least squares fitting we retrieve the electron return time of short trajectory by formulation in Kanai et al. [Phys. Rev. Lett. 98 (2007) 153904] when the gas jet is put after the laser focus.

Transient state of chiral recognition in a binary mixture of cyclodextrins in capillary electrophoresis

The transient state (as the defined point where no enantioseparation is obtained in a dual chiral selector system) of chiral recognition of aminoglutethimide in a binary mixture of neutral cyclodextrins (CDs) was studied by capillary electrophoresis (CE). The following three dual selector systems were used: alpha-cyclodextrin (alpha-CD) and beta-cyclodextrin (beta-CD); alpha-CD and heptakis(di-O-methyl-beta-cyclodextrin) (DM-beta-CD); alpha-CD and heptakis(tri-O-methyl-beta-cyclodextrin) (TM-beta-CD). The S-(-) enantiomer of the analyte was more strongly retained in the presence of either alpha-CD or TM-beta-CD at pH 2.5, 100 mM phosphate buffer, while the R-(+) enantiomer was more strongly retained in the presence of either P-CD or DM-P-CD. In the more simple case, the elution order is invariably kept if the enantiomers have the same elution order in either one of the two hosts of the binary mixture. In contrast, the elution order may be switched by varying the concentration ratio of two hosts that produce opposite elution order for this particular analyte. In such a dual selector system, the enantioselectivity will disappear at the transient state at a certain ratio of host,:host, Moreover, the migration times of the two enantiomers with host, alone (diluted in buffer) is approximately equal to the migration times at the corresponding concentration of host, alone (diluted in buffer), where the ratio of concentrations of host,:host, is the same as in the binary mixture at the transient state. As found by nuclear magnetic resonance experiments, the analyte is forming a 1:1 complex with either one of the CDs applied. From this finding, a theoretical model based on the mobility difference of the two enantiomers was derived that was used to simulate the transient state. (C) 2000 Elsevier Science B.V. All rights reserved.

Synthesis of Titania in Ethanol/Acetic Acid Mixture Solvents: Phase and Morphology Variations

The phase and morphology variations of titania prepared in ethanol/acetic acid mixture solvents have been systematically investigated. X-ray diffraction results and microscopy observations reveal that pure anatase aggregates consisted of small nanoparticles, pure rutile microspheres comprised of nanofibers, and their mixtures could be obtained by varying ratios of ethanol to acetic acid under solvothermal conditions. The contents of anatase and rutile in the mixed phases also vary with the ratios of ethanol to acetic acid. Field emission scanning electron microscopy and high resolution transmission electron microscopy results show that the two phases are separated from each other in final products and form aggregates with morphologies resembling to their pure phase products obtained under favorable conditions. The as-produced rutile nanofibers, either in pure phase or in mixed phases, tend to grow into hollow microspheres.

Theoretical representation of shear and pressure influence on the phase separation behavior of PVME/PS mixture

In the framework of lattice fluid model, the Gibbs energy and equation of state are derived by introducing the energy (E-s) stored during flow for polymer blends under shear. From the calculation of the spinodal of poly(vinyl methyl ether) (PVME) and polystyrene (PS) mixtures, we have found the influence of E., an equation of state in pure component is inappreciable, but it is appreciable in the mixture. However, the effect of E, on phase separation behavior is extremely striking. In the calculation of spinodal for the PVME/PS system, a thin, long and banana miscibility gap generated by shear is seen beside the miscibility gap with lower critical solution temperature. Meanwhile, a binodal coalescence of upper and lower miscibility gaps is occurred. The three points of the three-phase equilibrium are forecasted. The shear rate dependence of cloud point temperature at a certain composition is discussed. The calculated results are acceptable compared with the experiment values obtained by Higgins et at. However, the maximum positive shift and the minimum negative shift of cloud point temperature guessed by Higgins are not obtained, Furthermore, the combining effects of pressure and shear on spinodal shift are predicted.

Latest development of the combinatorial model of nuclear level densities

The combinatorial model of nuclear level densities has now reached a level of accuracy comparable to that of the best global analytical expressions without suffering from the limits imposed by the statistical hypothesis on which the latter expressions rely. In particular, it provides, naturally, non-Gaussian spin distribution as well as non-equipartition of parities which are known to have an impact on cross section predictions at low energies [1, 2, 3]. Our previous global models developed in Refs. [1, 2] suffered from deficiencies, in particular in the way the collective effects - both vibrational and rotational - were treated. We have recently improved this treatment using simultaneously the single-particle levels and collective properties predicted by a newly derived Gogny interaction [4], therefore enabling a microscopic description of energy-dependent shell, pairing and deformation effects. In addition for deformed nuclei, the transition to sphericity is coherently taken into account on the basis of a temperature-dependent Hartree-Fock calculation which provides at each temperature the structure properties needed to build the level densities. This new method is described and shown to give promising results with respect to available experimental data.

Speech Recognition with unknown partial feature corruption - a review of the union model

This paper provides a summary of our studies on robust speech recognition based on a new statistical approach – the probabilistic union model. We consider speech recognition given that part of the acoustic features may be corrupted by noise. The union model is a method for basing the recognition on the clean part of the features, thereby reducing the effect of the noise on recognition. To this end, the union model is similar to the missing feature method. However, the two methods achieve this end through different routes. The missing feature method usually requires the identity of the noisy data for noise removal, while the union model combines the local features based on the union of random events, to reduce the dependence of the model on information about the noise. We previously investigated the applications of the union model to speech recognition involving unknown partial corruption in frequency band, in time duration, and in feature streams. Additionally, a combination of the union model with conventional noise-reduction techniques was studied, as a means of dealing with a mixture of known or trainable noise and unknown unexpected noise. In this paper, a unified review, in the context of dealing with unknown partial feature corruption, is provided into each of these applications, giving the appropriate theory and implementation algorithms, along with an experimental evaluation.

On velocity-based local model networks for nonlinear identification

This paper exposes the strengths and weaknesses of the recently proposed velocity-based local model (LM) network. The global dynamics of the velocity-based blended representation are directly related to the dynamics of the underlying local models, an important property in the design of local controller networks. Furthermore, the sub-models are continuous-time and linear providing continuity with established linear theory and methods. This is not true for the conventional LM framework, where the global dynamics are only weakly related to the affine sub-models. In this paper, a velocity-based multiple model network is identified for a highly nonlinear dynamical system. The results show excellent dynamical modelling performances, highlighting the value of the velocity-based approach for the design and analysis of LM based control. Three important practical issues are also addressed. These relate to the blending of the velocity-based local models, the use of normalised Gaussian basis functions and the requirement of an input derivative.

Multiple model bootstrap filter for maneuvering target tracking

The extension of the bootstrap filter to the multiple model target tracking problem is considered. Bayesian bootstrap filtering is a very powerful technique since it represents samples by random samples and is therefore not restricted to linear, Gaussian systems, making it ideal for the multiple model problem where very complex densities fan be generated.

Artificial Intelligence Model for Flowable Concrete Mixtures Used in Underwater Construction and Repair

This study explores using artificial neural networks to predict the rheological and mechanical properties of underwater concrete (UWC) mixtures and to evaluate the sensitivity of such properties to variations in mixture ingredients. Artificial neural networks (ANN) mimic the structure and operation of biological neurons and have the unique ability of self-learning, mapping, and functional approximation. Details of the development of the proposed neural network model, its architecture, training, and validation are presented in this study. A database incorporating 175 UWC mixtures from nine different studies was developed to train and test the ANN model. The data are arranged in a patterned format. Each pattern contains an input vector that includes quantity values of the mixture variables influencing the behavior of UWC mixtures (that is, cement, silica fume, fly ash, slag, water, coarse and fine aggregates, and chemical admixtures) and a corresponding output vector that includes the rheological or mechanical property to be modeled. Results show that the ANN model thus developed is not only capable of accurately predicting the slump, slump-flow, washout resistance, and compressive strength of underwater concrete mixtures used in the training process, but it can also effectively predict the aforementioned properties for new mixtures designed within the practical range of the input parameters used in the training process with an absolute error of 4.6, 10.6, 10.6, and 4.4%, respectively.

TESTING FOR STATIONARITY IN HETEROGENEOUS PANEL DATA IN THE CASE OF MODEL MISSPECIFICATION

This paper investigates the performance of the tests proposed by Hadri and by Hadri and Larsson for testing for stationarity in heterogeneous panel data under model misspecification. The panel tests are based on the well known KPSS test (cf. Kwiatkowski et al.) which considers two models: stationarity around a deterministic level and stationarity around a deterministic trend. There is no study, as far as we know, on the statistical properties of the test when the wrong model is used. We also consider the case of the simultaneous presence of the two types of models in a panel. We employ two asymptotics: joint asymptotic, T, N -> infinity simultaneously, and T fixed and N allowed to grow indefinitely. We use Monte Carlo experiments to investigate the effects of misspecification in sample sizes usually used in practice. The results indicate that the assumption that T is fixed rather than asymptotic leads to tests that have less size distortions, particularly for relatively small T with large N panels (micro-panels) than the tests derived under the joint asymptotics. We also find that choosing a deterministic trend when a deterministic level is true does not significantly affect the properties of the test. But, choosing a deterministic level when a deterministic trend is true leads to extreme over-rejections. Therefore, when unsure about which model has generated the data, it is suggested to use the model with a trend. We also propose a new statistic for testing for stationarity in mixed panel data where the mixture is known. The performance of this new test is very good for both cases of T asymptotic and T fixed. The statistic for T asymptotic is slightly undersized when T is very small (

NON-GAUSSIAN STATISTICS OF OIL PRICING TIME-SERIES: A CASE STUDY

The stochastic nature of oil price fluctuations is investigated over a twelve-year period, borrowing feedback from an existing database (USA Energy Information Administration database, available online). We evaluate the scaling exponents of the fluctuations by employing different statistical analysis methods, namely rescaled range analysis (R/S), scale windowed variance analysis (SWV) and the generalized Hurst exponent (GH) method. Relying on the scaling exponents obtained, we apply a rescaling procedure to investigate the complex characteristics of the probability density functions (PDFs) dominating oil price fluctuations. It is found that PDFs exhibit scale invariance, and in fact collapse onto a single curve when increments are measured over microscales (typically less than 30 days). The time evolution of the distributions is well fitted by a Levy-type stable distribution. The relevance of a Levy distribution is made plausible by a simple model of nonlinear transfer. Our results also exhibit a degree of multifractality as the PDFs change and converge toward to a Gaussian distribution at the macroscales.