Bauer Method of MVDR Spectral Factorization for Pitch Modification in the Source Domain

In our earlier work [1], we employed MVDR (minimum variance distortionless response) based spectral estimation instead of modified-linear prediction method [2] in pitch modification. Here, we use the Bauer method of MVDR spectral factorization, leading to a causal inverse filter rather than a noncausal filter setup with MVDR spectral estimation [1]. Further, this is employed to obtain source (or residual) signal from pitch synchronous speech frames. The residual signal is resampled using DCT/IDCT depending on the target pitch scale factor. Finally, forward filters realized from the above factorization are used to get pitch modified speech. The modified speech is evaluated subjectively by 10 listeners and mean opinion scores (MOS) are tabulated. Further, modified bark spectral distortion measure is also computed for objective evaluation of performance. We find that the proposed algorithm performs better compared to time domain pitch synchronous overlap [3] and modified-LP method [2]. A good MOS score is achieved with the proposed algorithm compared to [1] with a causal inverse and forward filter setup.

A model based factorization approach for dense 3D recovery from monocular video

Feature track matrix factorization based methods have been attractive solutions to the Structure-front-motion (Sfnl) problem. Group motion of the feature points is analyzed to get the 3D information. It is well known that the factorization formulations give rise to rank deficient system of equations. Even when enough constraints exist, the extracted models are sparse due the unavailability of pixel level tracks. Pixel level tracking of 3D surfaces is a difficult problem, particularly when the surface has very little texture as in a human face. Only sparsely located feature points can be tracked and tracking error arc inevitable along rotating lose texture surfaces. However, the 3D models of an object class lie in a subspace of the set of all possible 3D models. We propose a novel solution to the Structure-from-motion problem which utilizes the high-resolution 3D obtained from range scanner to compute a basis for this desired subspace. Adding subspace constraints during factorization also facilitates removal of tracking noise which causes distortions outside the subspace. We demonstrate the effectiveness of our formulation by extracting dense 3D structure of a human face and comparing it with a well known Structure-front-motion algorithm due to Brand.

Design space exploration of systolic realization of QR factorization on a runtime reconfigurable platform

In the world of high performance computing huge efforts have been put to accelerate Numerical Linear Algebra (NLA) kernels like QR Decomposition (QRD) with the added advantage of reconfigurability and scalability. While popular custom hardware solution in form of systolic arrays can deliver high performance, they are not scalable, and hence not commercially viable. In this paper, we show how systolic solutions of QRD can be realized efficiently on REDEFINE, a scalable runtime reconfigurable hardware platform. We propose various enhancements to REDEFINE to meet the custom need of accelerating NLA kernels. We further do the design space exploration of the proposed solution for any arbitrary application of size n × n. We determine the right size of the sub-array in accordance with the optimal pipeline depth of the core execution units and the number of such units to be used per sub-array.

Hybrid online non-negative matrix factorization for clustering of documents

When document corpus is very large, we often need to reduce the number of features. But it is not possible to apply conventional Non-negative Matrix Factorization(NMF) on billion by million matrix as the matrix may not fit in memory. Here we present novel Online NMF algorithm. Using Online NMF, we reduced original high-dimensional space to low-dimensional space. Then we cluster all the documents in reduced dimension using k-means algorithm. We experimentally show that by processing small subsets of documents we will be able to achieve good performance. The method proposed outperforms existing algorithms.

Multivariate Granger causality: an estimation framework based on factorization of the spectral density matrix

Granger causality is increasingly being applied to multi-electrode neurophysiological and functional imaging data to characterize directional interactions between neurons and brain regions. For a multivariate dataset, one might be interested in different subsets of the recorded neurons or brain regions. According to the current estimation framework, for each subset, one conducts a separate autoregressive model fitting process, introducing the potential for unwanted variability and uncertainty. In this paper, we propose a multivariate framework for estimating Granger causality. It is based on spectral density matrix factorization and offers the advantage that the estimation of such a matrix needs to be done only once for the entire multivariate dataset. For any subset of recorded data, Granger causality can be calculated through factorizing the appropriate submatrix of the overall spectral density matrix.

Average Overlap for Clustering Incomplete Data using Symmetric Non-Negative Matrix Factorization

Clustering techniques which can handle incomplete data have become increasingly important due to varied applications in marketing research, medical diagnosis and survey data analysis. Existing techniques cope up with missing values either by using data modification/imputation or by partial distance computation, often unreliable depending on the number of features available. In this paper, we propose a novel approach for clustering data with missing values, which performs the task by Symmetric Non-Negative Matrix Factorization (SNMF) of a complete pair-wise similarity matrix, computed from the given incomplete data. To accomplish this, we define a novel similarity measure based on Average Overlap similarity metric which can effectively handle missing values without modification of data. Further, the similarity measure is more reliable than partial distances and inherently possesses the properties required to perform SNMF. The experimental evaluation on real world datasets demonstrates that the proposed approach is efficient, scalable and shows significantly better performance compared to the existing techniques.

Towards an exact factorization of the molecular wave function

An exact single-product factorisation of the molecular wave function for the timedependent Schrodinger equation is investigated by using an ansatz involving a phasefactor. By using the Frenkel variational method, we obtain the Schrodinger equations for the electronic and nuclear wave functions. The concept of a potential energy surface (PES) is retained by introducing a modified Hamiltonian as suggested earlier by Cederbaum. The parameter in the phase factor is chosen such that the equations of motion retain the physically appealing Born- Oppenheimer-like form, and is therefore unique.

Boxicity of Halin graphs

A k-dimensional box is the Cartesian product R-1 x R-2 x ... x R-k where each R-i is a closed interval on the real line. The boxicity of a graph G, denoted as box(G) is the minimum integer k such that G is the intersection graph of a collection of k-dimensional boxes. Halin graphs are the graphs formed by taking a tree with no degree 2 vertex and then connecting its leaves to form a cycle in such a way that the graph has a planar embedding. We prove that if G is a Halin graph that is not isomorphic to K-4, then box(G) = 2. In fact, we prove the stronger result that if G is a planar graph formed by connecting the leaves of any tree in a simple cycle, then box(G) = 2 unless G is isomorphic to K4 (in which case its boxicity is 1).

On the Conversion of Hensel Codes to Farey Rationals

Three different algorithms are described for the conversion of Hensel codes to Farey rationals. The first algorithm is based on the trial and error factorization of the weight of a Hensel code, inversion and range test. The second algorithm is deterministic and uses a pair of different p-adic systems for simultaneous computation; from the resulting weights of the two different Hensel codes of the same rational, two equivalence classes of rationals are generated using the respective primitive roots. The intersection of these two equivalence classes uniquely identifies the rational. Both the above algorithms are exponential (in time and/or space).

Spatial correlations around a Kondo impurity

We present a method for calculating zero-frequency, spatially varying response and correlation functions around a magnetic impurity in metals. We illustrate the procedure which involves a combination of perturbative scaling and nonperturbative renormalization-group methods for the spin-(1/2) Kondo impurity. We explain the factorization of the temperature and spatial dependence of the Knight shift observed in NMR experiments.

Total photoproduction cross section at very high energy

In this paper we apply to the photoproduction total cross section a model we have proposed for purely hadronic processes and which is based on QCD mini-jets and soft gluon re-summation. We compare the predictions of our model with the HERA data as well as with other models. For cosmic rays, our model predicts substantially higher cross sections at TeV energies than models based on factorization, but lower than models based on mini-jets alone, without soft gluons. We discuss the origin of this difference.

Linear time algorithms for Abelian group isomorphism and related problems

We consider the problem of determining if two finite groups are isomorphic. The groups are assumed to be represented by their multiplication tables. We present an O(n) algorithm that determines if two Abelian groups with n elements each are isomorphic. This improves upon the previous upper bound of O(n log n) [Narayan Vikas, An O(n) algorithm for Abelian p-group isomorphism and an O(n log n) algorithm for Abelian group isomorphism, J. Comput. System Sci. 53 (1996) 1-9] known for this problem. We solve a more general problem of computing the orders of all the elements of any group (not necessarily Abelian) of size n in O(n) time. Our algorithm for isomorphism testing of Abelian groups follows from this result. We use the property that our order finding algorithm works for any group to design a simple O(n) algorithm for testing whether a group of size n, described by its multiplication table, is nilpotent. We also give an O(n) algorithm for determining if a group of size n, described by its multiplication table, is Abelian. (C) 2007 Elsevier Inc. All rights reserved.

Hadwiger Number and the Cartesian Product of Graphs

The Hadwiger number eta(G) of a graph G is the largest integer n for which the complete graph K-n on n vertices is a minor of G. Hadwiger conjectured that for every graph G, eta(G) >= chi(G), where chi(G) is the chromatic number of G. In this paper, we study the Hadwiger number of the Cartesian product G square H of graphs. As the main result of this paper, we prove that eta(G(1) square G(2)) >= h root 1 (1 - o(1)) for any two graphs G(1) and G(2) with eta(G(1)) = h and eta(G(2)) = l. We show that the above lower bound is asymptotically best possible when h >= l. This asymptotically settles a question of Z. Miller (1978). As consequences of our main result, we show the following: 1. Let G be a connected graph. Let G = G(1) square G(2) square ... square G(k) be the ( unique) prime factorization of G. Then G satisfies Hadwiger's conjecture if k >= 2 log log chi(G) + c', where c' is a constant. This improves the 2 log chi(G) + 3 bound in [2] 2. Let G(1) and G(2) be two graphs such that chi(G1) >= chi(G2) >= clog(1.5)(chi(G(1))), where c is a constant. Then G1 square G2 satisfies Hadwiger's conjecture. 3. Hadwiger's conjecture is true for G(d) (Cartesian product of G taken d times) for every graph G and every d >= 2. This settles a question by Chandran and Sivadasan [2]. ( They had shown that the Hadiwger's conjecture is true for G(d) if d >= 3).

On Finding the Natural Number of Topics with Latent Dirichlet Allocation: Some Observations

It is important to identify the ``correct'' number of topics in mechanisms like Latent Dirichlet Allocation(LDA) as they determine the quality of features that are presented as features for classifiers like SVM. In this work we propose a measure to identify the correct number of topics and offer empirical evidence in its favor in terms of classification accuracy and the number of topics that are naturally present in the corpus. We show the merit of the measure by applying it on real-world as well as synthetic data sets(both text and images). In proposing this measure, we view LDA as a matrix factorization mechanism, wherein a given corpus C is split into two matrix factors M-1 and M-2 as given by C-d*w = M1(d*t) x Q(t*w).Where d is the number of documents present in the corpus anti w is the size of the vocabulary. The quality of the split depends on ``t'', the right number of topics chosen. The measure is computed in terms of symmetric KL-Divergence of salient distributions that are derived from these matrix factors. We observe that the divergence values are higher for non-optimal number of topics - this is shown by a `dip' at the right value for `t'.

Diffraction of a cylindrical pulse by a metallic half plane

A direct transform technique is applied to the initial and boundary value problem involving diffraction of a cylindrical pulse by a half plane, on which impedance type of boundary conditions must be met by the total field. The solution to the time harmonic incident plane wave is deduced as a particular case of the general time-dependent problem considered here and we avoid the Wiener–Hopf technique which leads to very complicated factorization and which masks the role of the impedance factor Z′ (a small quantity) in the expression for the scattered field.