On the relationship between hidden Markov Models and convex functional transforms for simulating scanning probe Microscopy

Models for simulating Scanning Probe Microscopy (SPM) may serve as a reference point for validating experimental data and practice. Generally, simulations use a microscopic model of the sample-probe interaction based on a first-principles approach, or a geometric model of macroscopic distortions due to the probe geometry. Examples of the latter include use of neural networks, the Legendre Transform, and dilation/erosion transforms from mathematical morphology. Dilation and the Legendre Transform fall within a general family of functional transforms, which distort a function by imposing a convex solution.In earlier work, the authors proposed a generalized approach to modeling SPM using a hidden Markov model, wherein both the sample-probe interaction and probe geometry may be taken into account. We present a discussion of the hidden Markov model and its relationship to these convex functional transforms for simulating and restoring SPM images.©2009 SPIE.

Constrained discriminative mapping transforms for unsupervised speaker adaptation

Discriminative mapping transforms (DMTs) is an approach to robustly adding discriminative training to unsupervised linear adaptation transforms. In unsupervised adaptation DMTs are more robust to unreliable transcriptions than directly estimating adaptation transforms in a discriminative fashion. They were previously proposed for use with MLLR transforms with the associated need to explicitly transform the model parameters. In this work the DMT is extended to CMLLR transforms. As these operate in the feature space, it is only necessary to apply a different linear transform at the front-end rather than modifying the model parameters. This is useful for rapidly changing speakers/environments. The performance of DMTs with CMLLR was evaluated on the WSJ 20k task. Experimental results show that DMTs based on constrained linear transforms yield 3% to 6% relative gain over MLE transforms in unsupervised speaker adaptation. © 2011 IEEE.

Geometrical properties and turbulent flame speed measurements in stationary premixed V-flames using direct numerical simulation

Three dimensional, fully compressible direct numerical simulations (DNS) of premixed turbulent flames are carried out in a V-flame configuration. The governing equations and the numerical implementation are described in detail, including modifications made to the Navier-Stokes Characteristic Boundary Conditions (NSCBC) to accommodate the steep transverse velocity and composition gradients generated when the flame crosses the boundary. Three cases, at turbulence intensities, u′/sL, of 1, 2, and 6 are considered. The influence of the flame holder on downstream flame properties is assessed through the distributions of the surface-conditioned displacement speed, curvature and tangential strain rates, and compared to data from similarly processed planar flames. The distributions are found to be indistinguishable from planar flames for distances greater than about 17δth downstream of the flame holder, where δth is the laminar flame thermal thickness. Favre mean fields are constructed, and the growth of the mean flame brush is found to be well described by simple Taylor type diffusion. The turbulent flame speed, sT is evaluated from an expression describing the propagation speed of an isosurface of the mean reaction progress variable c̃ in terms of the imbalance between the mean reactive, diffusive, and turbulent fluxes within the flame brush. The results are compared to the consumption speed, sC, calculated from the integral of the mean reaction rate, and to the predictions of a recently developed flame speed model (Kolla et al., Combust Sci Technol 181(3):518-535, 2009). The model predictions are improved in all cases by including the effects of mean molecular diffusion, and the overall agreement is good for the higher turbulence intensity cases once the tangential convective flux of c̃ is taken into account. © 2010 Springer Science+Business Media B.V.

Emerging modes of collective cell migration induced by geometrical constraints.

The role of geometrical confinement on collective cell migration has been recognized but has not been elucidated yet. Here, we show that the geometrical properties of the environment regulate the formation of collective cell migration patterns through cell-cell interactions. Using microfabrication techniques to allow epithelial cell sheets to migrate into strips whose width was varied from one up to several cell diameters, we identified the modes of collective migration in response to geometrical constraints. We observed that a decrease in the width of the strips is accompanied by an overall increase in the speed of the migrating cell sheet. Moreover, large-scale vortices over tens of cell lengths appeared in the wide strips whereas a contraction-elongation type of motion is observed in the narrow strips. Velocity fields and traction force signatures within the cellular population revealed migration modes with alternative pulling and/or pushing mechanisms that depend on extrinsic constraints. Force transmission through intercellular contacts plays a key role in this process because the disruption of cell-cell junctions abolishes directed collective migration and passive cell-cell adhesions tend to move the cells uniformly together independent of the geometry. Altogether, these findings not only demonstrate the existence of patterns of collective cell migration depending on external constraints but also provide a mechanical explanation for how large-scale interactions through cell-cell junctions can feed back to regulate the organization of migrating tissues.

From subspace learning to distance learning: A geometrical optimization approach

In this paper, we adopt a differential-geometry viewpoint to tackle the problem of learning a distance online. As this problem can be cast into the estimation of a fixed-rank positive semidefinite (PSD) matrix, we develop algorithms that exploits the rich geometry structure of the set of fixed-rank PSD matrices. We propose a method which separately updates the subspace of the matrix and its projection onto that subspace. A proper weighting of the two iterations enables to continuously interpolate between the problem of learning a subspace and learning a distance when the subspace is fixed. © 2009 IEEE.

Tandem system adaptation using multiple linear feature transforms

Adaptation to speaker and environment changes is an essential part of current automatic speech recognition (ASR) systems. In recent years the use of multi-layer percpetrons (MLPs) has become increasingly common in ASR systems. A standard approach to handling speaker differences when using MLPs is to apply a global speaker-specific constrained MLLR (CMLLR) transform to the features prior to training or using the MLP. This paper considers the situation when there are both speaker and channel, communication link, differences in the data. A more powerful transform, front-end CMLLR (FE-CMLLR), is applied to the inputs to the MLP to represent the channel differences. Though global, these FE-CMLLR transforms vary from time-instance to time-instance. Experiments on a channel distorted dialect Arabic conversational speech recognition task indicates the usefulness of adapting MLP features using both CMLLR and FE-CMLLR transforms. © 2013 IEEE.

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.

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.