Sparse Interpolation Research Articles

An Exemplar-Based Approach to Frequency Warping for Voice Conversion

The voice conversion's task is to modify a source speaker's voice to sound like that of a target speaker. A conversion method is considered successful when the produced speech sounds natural and similar to the target speaker. This paper presents a new voice conversion framework in which we combine frequency warping and exemplar-based method for voice conversion. Our method maintains high-resolution details during conversion by directly applying frequency warping on the high-resolution spectrum to represent the target. The warping function is generated by a sparse interpolation from a dictionary of exemplar warping functions. As the generated warping function is dependent only on a very small set of exemplars, we do away with the statistical averaging effects inherited from Gaussian mixture models. To compensate for the conversion error, we also apply residual exemplars into the conversion process. Both objective and subjective evaluations on the VOICES database validated the effectiveness of the proposed voice conversion framework. We observed a significant improvement in speech quality over the state-of-the-art parametric methods.

Reflectance and Natural Illumination from Single-Material Specular Objects Using Deep Learning.

In this paper, we present a method that estimates reflectance and illumination information from a single image depicting a single-material specular object from a given class under natural illumination. We follow a data-driven, learning-based approach trained on a very large dataset, but in contrast to earlier work we do not assume one or more components (shape, reflectance, or illumination) to be known. We propose a two-step approach, where we first estimate the object's reflectance map, and then further decompose it into reflectance and illumination. For the first step, we introduce a Convolutional Neural Network (CNN) that directly predicts a reflectance map from the input image itself, as well as an indirect scheme that uses additional supervision, first estimating surface orientation and afterwards inferring the reflectance map using a learning-based sparse data interpolation technique. For the second step, we suggest a CNN architecture to reconstruct both Phong reflectance parameters and high-resolution spherical illumination maps from the reflectance map. We also propose new datasets to train these CNNs. We demonstrate the effectiveness of our approach for both steps by extensive quantitative and qualitative evaluation in both synthetic and real data as well as through numerous applications, that show improvements over the state-of-the-art.

Sparse registration of diffusion weighted images

Background and objectiveRegistration is a critical step in group analysis of diffusion weighted images (DWI). Image registration is also necessary for construction of white matter atlases that can be used to identify white matter changes. A challenge in the registration of DWI is that the orientation of the fiber bundles should be considered in the process, making their registration more challenging than that of the scalar images. Most of the current registration methods use a model of diffusion profile, limiting the method to the used model. MethodsWe propose a model-independent method for DWI registration. The proposed method uses a multi-level free-form deformation (FFD), a sparse similarity measure, and a dictionary. We also propose a synthesis K-SVD algorithm for sparse interpolation of images during the registration process. We utilize two dictionaries: analysis dictionary is learned based on diffusion signals while synthesis dictionary is generated based on image patches. The proposed multi-level approach registers anatomical structures at different scales. T-test is used to determine the significance of the differences between different methods. ResultsWe have shown the efficiency of the proposed approach using real data. The method results in smaller generalized fractional anisotropy (GFA) root mean square (RMS) error (0.05 improvements, p = 0.0237) and angular error (0.37 ° improvement, p = 0.0330) compared to the large deformation diffeomorphic metric mapping (LDDMM) method and advanced normalization tools (ANTs). ConclusionSparse registration of diffusion signals enables registration of diffusion weighted images without using a diffusion model.

Multifidelity Sparse-Grid-Based Uncertainty Quantification for the Hokkaido Nansei-oki Tsunami

With uncertainty quantification, we aim to efficiently propagate the uncertainties in the input parameters of a computer simulation, in order to obtain a probability distribution of its output. In this work, we use multi-fidelity sparse grid interpolation to propagate the uncertainty in the shape of the incoming wave for the Okushiri test-case, which is a wave tank model of a part of the 1993 Hokkaido Nansei-oki tsunami. An important issue with many uncertainty quantification approaches is the ‘curse of dimensionality’: the overall computational cost of the uncertainty propagation increases rapidly when we increase the number of uncertain input parameters. We aim to mitigate the curse of dimensionality by using a multifidelity approach. In the multifidelity approach, we combine results from a small number of accurate and expensive high-fidelity simulations with a large number of less accurate but also less expensive low-fidelity simulations. For the Okushiri test-case, we find an improved scaling when we increase the number of uncertain input parameters. This results in a significant reduction of the overall computational cost. For example, for four uncertain input parameters, accurate uncertainty quantification based on only high-fidelity simulations comes at a normalised cost of 219 high-fidelity simulations; when we use a multifidelity approach, this is reduced to a normalised cost of only 10 high-fidelity simulations.

Practical application of collocation method in power flow study of South Australia grid

In this paper stochastic collocation method is proposed to solve probabilistic power flow (PPF) model of South Australia (SA). And this model is based upon historical acquisition of power system data of SA. In SA, numbers of wind farms are installed and the variability of wind speed brings more uncertainties into the power system. The traditional deterministic power flow (DPF) computation does not take into account the probabilistic nature of power system uncertainties, so PPF computation is imperative. However, as a commonly used PPF simulation method, Monte Carlo simulation (MCS) has a very high computation cost. Hence, in this paper sparse grid interpolation (SGI) is presented to accomplish PPF analysis with striking high computation efficiency. Meanwhile, instead of using theoretical wind power generation model, probabilistic collocation method (PCM) is used to construct realistic relationship between wind speed and wind power. In addition, fuzzy logic optimization is applied to PCM to improve the accuracy of the model output. The paper concludes with presentation of an aggregated DC power flow model of SA to compare the computation efficiency of the SGI and MC.

The nonparametric Bayesian dictionary learning based interpolation method for WSNs missing data

The conventional data interpolation methods based on sparse representation usually assume that the signal is sparse under the overcomplete dictionary. Specially, they must confirm the dimensions of dictionary and the signal sparse level in advance. However, it is hard to know them if the signal is complicated or dynamically changing. In this paper, we proposed a nonparametric Bayesian dictionary learning based interpolation method for WSNs missing data, which is the combination of sparse representation and data interpolation. This method need not preset sparse degrees and dictionary dimensions, and our dictionary atoms are drawn from a multivariate normal distribution. In this case, the dictionary size will be learned adaptively by the nonparametric Bayesian method. In addition, we implement the Dirichlet process to exploit the spatial similarity of the sensing data in WSNs, thus to improve the interpolation accuracy. The interpolation model parameters, the optimal dictionary and sparse coefficients, can be inferred by the means of Gibbs sampling. The missing data will be estimated commendably through the derived parameters. The experimental results show that the data interpolation method we proposed outperforms the conventional methods in terms of interpolation accuracy and robustness.

Multilevel quasi-interpolation on a sparse grid with the Gaussian

Motivated by the recent multilevel sparse kernel-based interpolation (MuSIK) algorithm proposed in Georgoulis et al. (SIAM J. Sci. Comput. 35, 815–832, 2013), we introduce the new quasi-multilevel sparse interpolation with kernels (Q-MuSIK) via the combination technique. The Q-MuSIK scheme achieves better convergence and run time when compared with classical quasi-interpolation. Also, the Q-MuSIK algorithm is generally superior to the MuSIK methods in terms of run time in particular in high-dimensional interpolation problems, since there is no need to solve large algebraic systems. We subsequently propose a fast, low complexity, high-dimensional positive-weight quadrature formula based on Q-MuSIKSapproximation of the integrand. We present the results of numerical experimentation for both quasi-interpolation and quadrature in high dimensions.

Towards non-intrusive reduced order 3D free surface flow modelling

In this article, we describe a novel non-intrusive reduction model for three-dimensional (3D) free surface flows. However, in this work we limit the vertical resolution to be a single element. So, although it does resolve some non-hydrostatic effects, it does not examine the application of reduced modelling to full 3D free surface flows, but it is an important step towards 3D modelling. A newly developed non-intrusive reduced order model (NIROM) (Xiao et al., 2015a) has been used in this work. Rather than taking the standard POD approach using the Galerkin projection, a Smolyak sparse grid interpolation method is employed to generate the NIROM. A set of interpolation functions is constructed to calculate the POD coefficients, where the POD coefficients at previous time steps are the inputs of the interpolation function. Therefore, this model is non-intrusive and does not require modifications to the code of the full system and is easy to implement.By using this new NIROM, we have developed a robust and efficient reduced order model for free surface flows within a 3D unstructured mesh finite element ocean model. What distinguishes the reduced order model developed here from other existing reduced order ocean models is (1) the inclusion of 3D dynamics with a free surface (the 3D computational domain and meshes are changed with the movement of the free surface); (2) the incorporation of wetting-drying; and (3) the first implementation of non-intrusive reduced order method in ocean modelling. Most importantly, the change of the computational domain with the free surface movement is taken into account in reduced order modelling. The accuracy and predictive capability of the new non-intrusive free surface flow ROM have been evaluated in Balzano and Okushiri tsunami test cases. This is the first step towards 3D reduced order modelling in realistic ocean cases. Results obtained show that the accuracy of free surface problems relative to the high fidelity model is maintained in ROM whilst the CPU time is reduced by several orders of magnitude.

An algorithm for symbolic-numeric sparse interpolation of multivariate polynomials whose degree bounds are unknown

We consider the problem of sparse interpolation of a multivariate black-box polynomial in floatingpoint arithmetic. More specifically, we assume that we are given a black-box polynomial f ( x 1 ,... x n ) = Σ t j =1 c j x 1 d j , 1 ... x n d j , n ∈ C[ x 1 ,..., x n ] ( c j ≠ 0)and the number of terms t , and that we can evaluate the value of f ( x 1 ,...,x n ) at any point in C n in floating-point arithmetic. The problem is to find the coefficients c 1 , ..., c t and the exponents d 1,1, ..., d t,n . We propose an efficient algorithm to solve the problem.

An Improved Early Termination Sparse Interpolation Algorithm for Multivariate Polynomials

This paper presents an improved early termination algorithm for sparse black box multivariate polynomials, which reduces the interpolation problem into several sub-interpolation problems with less variables and fewer terms. Actually, all interpolations are eventually reduced to the interpolation of a list of polynomials with less terms than that of the original polynomial. Extensive experiments show that the new algorithm is much faster than the original algorithm.

Canny edge-based deformable image registration

This work focuses on developing a 2D Canny edge-based deformable image registration (Canny DIR) algorithm to register in vivo white light images taken at various time points. This method uses a sparse interpolation deformation algorithm to sparsely register regions of the image with strong edge information. A stability criterion is enforced which removes regions of edges that do not deform in a smooth uniform manner. Using a synthetic mouse surface ground truth model, the accuracy of the Canny DIR algorithm was evaluated under axial rotation in the presence of deformation. The accuracy was also tested using fluorescent dye injections, which were then used for gamma analysis to establish a second ground truth. The results indicate that the Canny DIR algorithm performs better than rigid registration, intensity corrected Demons, and distinctive features for all evaluation matrices and ground truth scenarios. In conclusion Canny DIR performs well in the presence of the unique lighting and shading variations associated with white-light-based image registration.

Plug-and-Play Priors for Bright Field Electron Tomography and Sparse Interpolation

Many material and biological samples in scientific imaging are characterized by non-local repeating structures. These are studied using scanning electron microscopy and electron tomography. Sparse sampling of individual pixels in a 2D image acquisition geometry, or sparse sampling of projection images with large tilt increments in a tomography experiment, can enable high speed data acquisition and minimize sample damage caused by the electron beam. In this paper, we present an algorithm for electron tomographic reconstruction and sparse image interpolation that exploits the non-local redundancy in images. We adapt a framework, termed plug-and-play (P&P) priors, to solve these imaging problems in a regularized inversion setting. The power of the P&P approach is that it allows a wide array of modern denoising algorithms to be used as a "prior model" for tomography and image interpolation. We also present sufficient mathematical conditions that ensure convergence of the P&P approach, and we use these insights to design a new non-local means denoising algorithm. Finally, we demonstrate that the algorithm produces higher quality reconstructions on both simulated and real electron microscope data, along with improved convergence properties compared to other methods.

Sparse polynomials in FLINT

We have implemented a high-performance C library for sparse polynomials, which is provided as an add-on module to the open-source computation library FLINT [7]. Our implementation incorporates a number of recent theoretical advances in supersparse polynomial arithmetic, most notably recent algorithms for sparse interpolation and multiplication. We provide a summary of the provided functionality, a selection of key implementation decisions, and some preliminary timing data.

A two-stage adaptive stochastic collocation method on nested sparse grids for multiphase flow in randomly heterogeneous porous media

A new computational method is proposed for efficient uncertainty quantification of multiphase flow in porous media with stochastic permeability. For pressure estimation, it combines the dimension-adaptive stochastic collocation method on Smolyak sparse grids and the Kronrod–Patterson–Hermite nested quadrature formulas. For saturation estimation, an additional stage is developed, in which the pressure and velocity samples are first generated by the sparse grid interpolation and then substituted into the transport equation to solve for the saturation samples, to address the low regularity problem of the saturation. Numerical examples are presented for multiphase flow with stochastic permeability fields to demonstrate accuracy and efficiency of the proposed two-stage adaptive stochastic collocation method on nested sparse grids.

Precomputing strategy for Hamiltonian Monte Carlo method based on regularity in parameter space

Markov Chain Monte Carlo (MCMC) algorithms play an important role in statistical inference problems dealing with intractable probability distributions. Recently, many MCMC algorithms such as Hamiltonian Monte Carlo (HMC) and Riemannian Manifold HMC have been proposed to provide distant proposals with high acceptance rate. These algorithms, however, tend to be computationally intensive which could limit their usefulness, especially for big data problems due to repetitive evaluations of functions and statistical quantities that depend on the data. This issue occurs in many statistic computing problems. In this paper, we propose a novel strategy that exploits smoothness (regularity) of parameter space to improve computational efficiency of MCMC algorithms. When evaluation of functions or statistical quantities are needed at a point in parameter space, interpolation from precomputed values or previous computed values is used. More specifically, we focus on Hamiltonian Monte Carlo (HMC) algorithms that use geometric information for faster exploration of probability distributions. Our proposed method is based on precomputing the required geometric information on a set of grids before running sampling information at nearby grids at each iteration of HMC. Sparse grid interpolation method is used for high dimensional problems. Tests on computational examples are shown to illustrate the advantages of our method.

Sparse multivariate polynomial interpolation on the basis of Schubert polynomials

Schubert polynomials were discovered by A. Lascoux and M. Sch\"utzenberger in the study of cohomology rings of flag manifolds in 1980's. These polynomials generalize Schur polynomials, and form a linear basis of multivariate polynomials. In 2003, Lenart and Sottile introduced skew Schubert polynomials, which generalize skew Schur polynomials, and expand in the Schubert basis with the generalized Littlewood-Richardson coefficients. In this paper we initiate the study of these two families of polynomials from the perspective of computational complexity theory. We first observe that skew Schubert polynomials, and therefore Schubert polynomials, are in $\CountP$ (when evaluating on non-negative integral inputs) and $\VNP$. Our main result is a deterministic algorithm that computes the expansion of a polynomial $f$ of degree $d$ in $\Z[x_1, \dots, x_n]$ in the basis of Schubert polynomials, assuming an oracle computing Schubert polynomials. This algorithm runs in time polynomial in $n$, $d$, and the bit size of the expansion. This generalizes, and derandomizes, the sparse interpolation algorithm of symmetric polynomials in the Schur basis by Barvinok and Fomin (Advances in Applied Mathematics, 18(3):271--285). In fact, our interpolation algorithm is general enough to accommodate any linear basis satisfying certain natural properties. Applications of the above results include a new algorithm that computes the generalized Littlewood-Richardson coefficients.

Library-Based Sparse Interpolation and Super-Resolution of S/TEM Images of Biological and Material Nano-Structures

Library-Based Sparse Interpolation and Super-Resolution of S/TEM Images of Biological and Material Nano-Structures

Efficient approximation of Sparse Jacobians for time‐implicit reduced order models

SummaryThis paper introduces a sparse matrix discrete interpolation method to effectively compute matrix approximations in the reduced order modeling framework. The sparse algorithm developed herein relies on the discrete empirical interpolation method and uses only samples of the nonzero entries of the matrix series. The proposed approach can approximate very large matrices, unlike the current matrix discrete empirical interpolation method, which is limited by its large computational memory requirements. The empirical interpolation indices obtained by the sparse algorithm slightly differ from the ones computed by the matrix discrete empirical interpolation method as a consequence of the singular vectors round‐off errors introduced by the economy or full singular value decomposition (SVD) algorithms when applied to the full matrix snapshots. When appropriately padded with zeros, the economy SVD factorization of the nonzero elements of the snapshots matrix is a valid economy SVD for the full snapshots matrix. Numerical experiments are performed with the 1D Burgers and 2D shallow water equations test problems where the quadratic reduced nonlinearities are computed via tensorial calculus. The sparse matrix approximation strategy is compared against five existing methods for computing reduced Jacobians: (i) matrix discrete empirical interpolation method, (ii) discrete empirical interpolation method, (iii) tensorial calculus, (iv) full Jacobian projection onto the reduced basis subspace, and (v) directional derivatives of the model along the reduced basis functions. The sparse matrix method outperforms all other algorithms. The use of traditional matrix discrete empirical interpolation method is not possible for very large dimensions because of its excessive memory requirements. Copyright © 2016 John Wiley & Sons, Ltd.

Dimension-Adaptive Sparse Grid Interpolation for Uncertainty Quantification in Modern Power Systems: Probabilistic Power Flow

In this paper, the authors firstly present the theoretical foundation of a state-of-the-art uncertainty quantification method, the dimension-adaptive sparse grid interpolation (DASGI), for introducing it into the applications of probabilistic power flow (PPF), specifically as discussed herein. It is well-known that numerous sources of uncertainty are being brought into the present-day electrical grid, by large-scale integration of renewable, thus volatile, generation, e.g., wind power, and by unprecedented load behaviors. In presence of these added uncertainties, it is imperative to change traditional deterministic power flow (DPF) calculation to take them into account in the routine operation and planning. However, the PPF analysis is still quite challenging due to two features of the uncertainty in modern power systems: high dimensionality and presence of stochastic interdependence. Both are traditionally addressed by the Monte Carlo simulation (MCS) at the cost of cumbersome computation; in this paper instead, they are tackled with the joint application of the DASGI and Copula theory (especially advantageous for constructing nonlinear dependence among various uncertainty sources), in order to accomplish the dependent high-dimensional PPF analysis in an accurate and faster manner. Based on the theory of DASGI, its combination with Copula and the DPF for the PPF is also introduced systematically in this work. Finally, the feasibility and the effectiveness of this methodology are validated by the test results of two standard IEEE test cases.

Polynomial Homotopy Method for the Sparse Interpolation Problem Part I: Equally Spaced Sampling

In this paper, the solution of the polynomial system arising from sparse interpolation problems with equally spaced sampling is studied. Exploiting the special structure of the polynomial system, it is proved that all its solutions lie in one orbit for generic data. Then, based on the coefficient parameter homotopy method, an efficient algorithm for solving the system is proposed. Smoothness and accessibility of the coefficient parameter homotopy hold for the constant terms of the start system generated by a randomly chosen starting point. So, in the coefficient parameter homotopy method, it needs no computational cost in the first phase and only one path needs to be traced in the second phase. Preliminary numerical tests show that the proposed algorithm is promising.