An efficient sampling method for stochastic inverse problems

A general framework is developed to treat inverse problems with parameters that are random fields. It involves a sampling method that exploits the sensitivity derivatives of the control variable with respect to the random parameters. As the sensitivity derivatives are computed only at the mean values of the relevant parameters, the related extra cost of the present method is a fraction of the total cost of the Monte Carlo method. The effectiveness of the method is demonstrated on an example problem governed by the Burgers equation with random viscosity. It is specifically shown that this method is two orders of magnitude more efficient compared to the conventional Monte Carlo method. In other words, for a given number of samples, the present method yields two orders of magnitude higher accuracy than its conventional counterpart.     

On the covariance of the level sizes in random recursive trees

In this paper we study the covariance structure of the number of nodes k and l steps away from the root in random recursive trees. We give an analytic expression valid for all k, l and tree sizes N. The fraction of nodes k steps away from the root is a random probability distribution in k. The expression for the covariances allows us to show that the total variation distance between this (random) probability distribution and its mean converges in probability to zero. © 2002 Wiley Periodicals, Inc. Random Struct. Alg., 20: 519–539, 2002     

Fastiterative methods for least squares estimations

Least squares estimations have been used extensively in many applications, e.g. system identification and signal prediction. When the stochastic process is stationary, the least squares estimators can be found by solving a Toeplitz or near-Toeplitz matrix system depending on the knowledge of the data statistics. In this paper, we employ the preconditioned conjugate gradient method with circulant preconditioners to solve such systems. Our proposed circulant preconditioners are derived from the spectral property of the given stationary process. In the case where the spectral density functions() of the process is known, we prove that ifs() is a positive continuous function, then the spectrum of the preconditioned system will be clustered around 1 and the method converges superlinearly. However, if the statistics of the process is unknown, then we prove that with probability 1, the spectrum of the preconditioned system is still clustered around 1 provided that large data samples are taken. For finite impulse response (FIR) system identification problems, our numerical results show that annth order least squares estimator can usually be obtained inO(n logn) operations whenO(n) data samples are used. Finally, we remark that our algorithm can be modified to suit the applications of recursive least squares computations with the proper use of sliding window method arising in signal processing applications.Research supported in part by HKRGC grant no. 221600070, ONR contract no. N00014-90-J-1695 and DOE grant no. DE-FG03-87ER25037.     

The covariance matrix of the Potts model: A random cluster analysis

We consider the covariance matrix,G ^mm =q ²<(_x,m);(_y,m)>, of thed-dimensionalq-states Potts model, rewriting it in the random cluster representation of Fortuin and Kasteleyn. In any of theq ordered phases, we identify the eigenvalues of this matrix both in terms of representations of the unbroken symmetry group of the model and in terms of random cluster connectivities and covariances, thereby attributing algebraic significance to these stochastic geometric quantities. We also show that the correlation length corresponding to the decay rate of one of the eigenvalues is the same as the inverse decay rate of the diameter of finite clusers. For dimensiond=2, we show that this correlation length and the correlation length of the two-point function with free boundary conditions at the corresponding dual temperature are equal up to a factor of two. For systems with first-order transitions, this relation helps to resolve certain inconsistencies between recent exact and numerical work on correlation lengths at the self-dual point _o. For systems with second order transitions, this relation implies the equality of the correlation length exponents from above and below threshold, as well as an amplitude ratio of two. In the course of proving the above results, we establish several properties of independent interest, including left continuity of the inverse correlation length with free boundary conditions and upper semicontinuity of the decay rate for finite clusters in all dimensions, and left continuity of the two-dimensional free boundary condition percolation probability at _o. We also introduce DLR equations for the random cluster model and use them to establish ergodicity of the free measure. In order to prove these results, we introduce a new class of events which we call decoupling events and two inequalities for these events. The first is similar to the FKG inequality, but holds for events which are neither increasing nor decreasing; the second is similar to the van den Berg-Kesten inequality in standard percolation. Both inequalities hold for an arbitrary FKG measure.     

Dimer asymmetry in superoxide dismutase studied by molecular dynamics simulation

Summary Molecular dynamics (MD) simulations of 100 ps have been carried out to study the active-site behaviour of the Cu,Zn superoxide dismutase dimer (SOD) in water. The active site of each subunit was monitored during the whole simulation by calculating the distances between functional residues and the catalytic copper. The results indicate that charge orientation is maintained at each active site but the solvent accessibility varies. Analysis of the MD simulation, carried out by using the atomic displacement covariance matrix, has shown a different intra-subunit correlation pattern for the two monomers and the presence of inter-subunit correlations. The MD simulation presented here indicates an asymmetry in the two active sites and different dynamic behaviour of the two SOD subunits.     

Limit theorems for the pre-averaged Hayashi–Yoshida estimator with random sampling

We will focus on estimating the integrated covariance of two diffusion processes observed in a nonsynchronous manner. The observation data is contaminated by some noise, which possibly depends on the time and the latent diffusion processes, while the sampling times also possibly depend on the observed processes. In a high-frequency setting, we consider a modified version of the pre-averaged Hayashi–Yoshida estimator, and we show that such a kind of estimator has the consistency and the asymptotic mixed normality, and attains the optimal rate of convergence.     

Liquid crystal mixtures for potential infrared applications

Liquid crystals (LCs) with diazo linkages have high dielectric and optical anisotropy. Two newly synthesized liquid crystalline compounds were dissolved in room temperature LC hosts, E7 and PTTP-24/36, to assess their properties. It was found that these mixtures have higher birefringence, larger dielectric anisotropy, and better elastic properties than the hosts. Satisfactory viscoelastic coefficients were also obtained for these mixtures, showing that they are promising LC materials for applications in the near IR region.     

The convergence on spectrum of sample covariance matrices for information-plus-noise type data

In this paper,we consider the limiting spectral distribution of the information-plusnoise type sample covariance matrices Cn =1/N (Rn + σXn) (Rn + σXn)*,under the assumption that the entries of Xn are ...     

Sequential paired covariance for improved visualization of mass spectrometry imaging datasets

Untargeted analyses in mass spectrometry imaging produce hundreds of ion images representing spatial distributions of biomolecules in biological tissues. Due to the large diversity of ions detected in untargeted analyses, normalization standards are often difficult to implement to account for pixel-to-pixel variability in imaging studies. Many normalization strategies exist to account for this variability, but they largely do not improve image quality. In this study, we present a new approach for improving image quality and visualization of tissue features by application of sequential paired covariance (SPC). This approach was demonstrated using previously published tissue datasets such as rat brain and human prostate with different biomolecules like metabolites and N-linked glycans. Data transformation by SPC improved ion images resulting in increased smoothing of biological features compared with commonly used normalization approaches.     

Chimera Spectrum Diagnostics for Peptides Using Two-Dimensional Partial Covariance Mass Spectrometry

The rate of successful identification of peptide sequences by tandem mass spectrometry (MS/MS) is adversely affected by the common occurrence of co-isolation and co-fragmentation of two or more isobaric or isomeric parent ions. This results in so-called `chimera spectra’, which feature peaks of the fragment ions from more than a single precursor ion. The totality of the fragment ion peaks in chimera spectra cannot be assigned to a single peptide sequence, which contradicts a fundamental assumption of the standard automated MS/MS spectra analysis tools, such as protein database search engines. This calls for a diagnostic method able to identify chimera spectra to single out the cases where this assumption is not valid. Here, we demonstrate that, within the recently developed two-dimensional partial covariance mass spectrometry (2D-PC-MS), it is possible to reliably identify chimera spectra directly from the two-dimensional fragment ion spectrum, irrespective of whether the co-isolated peptide ions are isobaric up to a finite mass accuracy or isomeric. We introduce ‘3-57 chimera tag’ technique for chimera spectrum diagnostics based on 2D-PC-MS and perform numerical simulations to examine its efficiency. We experimentally demonstrate the detection of a mixture of two isomeric parent ions, even under conditions when one isomeric peptide is at one five-hundredth of the molar concentration of the second isomer.