Weaknesses of Margulis and Ramanujan-Margulis low-density parity-check codes

We report weaknesses in two algebraic constructions of low-density parity-check codes based on expander graphs. The Margulis construction gives a code with near-codewords, which cause problems for the sum-product decoder; The Ramanujan-Margulis construction gives a code with low-weight codewords, which produce an error-floor. © 2004 Elsevier B.V.

Weaknesses of margulis and ramanujan-margulis low-density parity-check codes

We report weaknesses in two algebraic constructions of low-density parity-check codes based on expander graphs. The Margulis construction gives a code with near-codewords, which cause problems for the sum-product decoder; The Ramanujan-Margulis construction gives a code with low-weight codewords, which produce an error-floor. ©2003 Published by Elsevier Science B. V.

Development of low entropy coding in a recurrent network.

In this paper we present an unsupervised neural network which exhibits competition between units via inhibitory feedback. The operation is such as to minimize reconstruction error, both for individual patterns, and over the entire training set. A key difference from networks which perform principal components analysis, or one of its variants, is the ability to converge to non-orthogonal weight values. We discuss the network's operation in relation to the twin goals of maximizing information transfer and minimizing code entropy, and show how the assignment of prior probabilities to network outputs can help to reduce entropy. We present results from two binary coding problems, and from experiments with image coding.

The rise heights of low-and high-Froude-number turbulent axisymmetric fountains

We present the results of an experimental investigation across a broad range of source Froude numbers, 0. 4 ≤ Fr 0 ≤ 45, into the dynamics, morphology and rise heights of Boussinesq turbulent axisymmetric fountains in quiescent uniform environments. Typically, these fountains are thought to rise to an initial height, z i, before settling back and fluctuating about a lesser (quasi-) steady height, z ss. Our measurements show that this is not always the case and the ratio of the fountain's initial rise height to steady rise height, λ = z i/z ss, varies widely, 0. 5 ≈ λ ≈ 2, across the range of Fr 0 investigated. As a result of near-ideal start-up conditions provided by the experimental set-up we were consistently able to form a vortex at the fountain's front. This enabled new insights into two features of the initial rise of turbulent fountains. Firstly, for 1. 0 ≈ Fr 0 ≈ 1. 7 the initial rise height is less than the steady rise height. Secondly, for Fr 0 ≈ 5. 5, the vortex formed at the fountain's front pinches off, separates from the main body and rises high above the fountain; there is thus a third rise height to consider, namely, the maximum vortex rise height, z v. From our observations we propose classifying turbulent axisymmetric fountains into five regimes (as opposed to the current three regimes) and present detailed descriptions of the flow in each. Finally, based on an analysis of the rise height fluctuations and the width of fountains in (quasi-) steady state we provide further insight into the physical cause of height fluctuations. © 2011 Cambridge University Press.

Low-rank optimization for distance matrix completion

This paper addresses the problem of low-rank distance matrix completion. This problem amounts to recover the missing entries of a distance matrix when the dimension of the data embedding space is possibly unknown but small compared to the number of considered data points. The focus is on high-dimensional problems. We recast the considered problem into an optimization problem over the set of low-rank positive semidefinite matrices and propose two efficient algorithms for low-rank distance matrix completion. In addition, we propose a strategy to determine the dimension of the embedding space. The resulting algorithms scale to high-dimensional problems and monotonically converge to a global solution of the problem. Finally, numerical experiments illustrate the good performance of the proposed algorithms on benchmarks. © 2011 IEEE.

Low-rank optimization on the cone of positive semidefinite matrices

We propose an algorithm for solving optimization problems defined on a subset of the cone of symmetric positive semidefinite matrices. This algorithm relies on the factorization X = Y Y T , where the number of columns of Y fixes an upper bound on the rank of the positive semidefinite matrix X. It is thus very effective for solving problems that have a low-rank solution. The factorization X = Y Y T leads to a reformulation of the original problem as an optimization on a particular quotient manifold. The present paper discusses the geometry of that manifold and derives a second-order optimization method with guaranteed quadratic convergence. It furthermore provides some conditions on the rank of the factorization to ensure equivalence with the original problem. In contrast to existing methods, the proposed algorithm converges monotonically to the sought solution. Its numerical efficiency is evaluated on two applications: the maximal cut of a graph and the problem of sparse principal component analysis. © 2010 Society for Industrial and Applied Mathematics.

Acoustic optimization of ultra-low count bypass outlet guide vanes

An optimization process has been used to design an ultra-low count fan outlet guide vane with an unconventional leading edge profile to reduce the interaction noise. Computational fluid dynamics has been used to predict the aerodynamic and acoustic performance of the stator vane. The final stator design has been built and tested in a representative fan stage rig to determine its tone noise characteristics. The stator vane is found to give significant tone noise reduction at the fundamental blade passing frequency at cut-back in line with design expectations. Detailed comparisons of predicted circumferential and radial modes levels against measured mode detection data are also presented. A good agreement was found between numerical predictions and experimental data.

Low-rank optimization with trace norm penalty

The paper addresses the problem of low-rank trace norm minimization. We propose an algorithm that alternates between fixed-rank optimization and rank-one updates. The fixed-rank optimization is characterized by an efficient factorization that makes the trace norm differentiable in the search space and the computation of duality gap numerically tractable. The search space is nonlinear but is equipped with a Riemannian structure that leads to efficient computations. We present a second-order trust-region algorithm with a guaranteed quadratic rate of convergence. Overall, the proposed optimization scheme converges superlinearly to the global solution while maintaining complexity that is linear in the number of rows and columns of the matrix. To compute a set of solutions efficiently for a grid of regularization parameters we propose a predictor-corrector approach that outperforms the naive warm-restart approach on the fixed-rank quotient manifold. The performance of the proposed algorithm is illustrated on problems of low-rank matrix completion and multivariate linear regression. © 2013 Society for Industrial and Applied Mathematics.

Low voltage ride-through of DFIG and brushless DFIG: Similarities and differences

The brushless doubly fed induction generator (BDFIG) has been proposed as a viable alternative in wind turbines to the commonly used doubly fed induction generator (DFIG). The BDFIG retains the benefits of the DFIG, i.e. variable speed operation with a partially rated converter, but without the use of brush gear and slip rings, thereby conferring enhanced reliability. As low voltage ride-through (LVRT) performance of the DFIG-based wind turbine is well understood, this paper aims to analyze LVRT behavior of the BDFIG-based wind turbine in a similar way. In order to achieve this goal, the equivalence between their two-axis model parameters is investigated. The variation of flux linkages, back-EMFs and currents of both types of generator are elaborated during three phase voltage dips. Moreover, the structural differences between the two generators, which lead to different equivalent parameters and hence different LVRT capabilities, are investigated. The analytical results are verified via time-domain simulations for medium size wind turbine generators as well as experimental results of a voltage dip on a prototype 250 kVA BDFIG. © 2014 Elsevier B.V.

The low velocity impact response of sandwich beams with a corrugated core or a Y-frame core

The low speed impact responses of simply-supported and clamped sandwich beams with corrugated and Y-frame cores have been measured in a drop-weight apparatus at 5 m s-1. The AISI 304 stainless steel sandwich beams comprised two identical face sheets and represented 1:20 scale versions of ship hull designs. No significant rate effects were observed at impact speeds representative of ship collisions: the drop-weight responses were comparable to the ones measured quasi-statically. Moreover, the corrugated and Y-frame core beams had similar performances. Three-dimensional finite element (FE) models simulated the experiments and were in good agreement with the measurements. The simulations demonstrated correctly that the sandwich beams collapsed by core indentation under both quasi-static loading and in the drop-weight experiments. These FE models were then used to investigate the sensitivity of impact response to (i) velocity, over a wider range of velocities than achievable with the drop-weight apparatus, and (ii) the presence of the back face sheet. The dynamic responses of sandwich beams with both front and back face sheets were found to be within 20% of the quasi-static responses for speeds less than approximately 5 m s-1. This suggests that quasi-static considerations are adequate to model the collision of a sandwich ship hull. By contrast, beams without a back face collapsed by Brazier buckling under quasi-static loading conditions, and by core indentation at a loading velocity of 5 m s-1. Thus, dynamic considerations are needed in ship hull designs that do not employ a back face. © 2014 Elsevier Ltd. All rights reserved.

Fixed-rank matrix factorizations and Riemannian low-rank optimization

Motivated by the problem of learning a linear regression model whose parameter is a large fixed-rank non-symmetric matrix, we consider the optimization of a smooth cost function defined on the set of fixed-rank matrices. We adopt the geometric framework of optimization on Riemannian quotient manifolds. We study the underlying geometries of several well-known fixed-rank matrix factorizations and then exploit the Riemannian quotient geometry of the search space in the design of a class of gradient descent and trust-region algorithms. The proposed algorithms generalize our previous results on fixed-rank symmetric positive semidefinite matrices, apply to a broad range of applications, scale to high-dimensional problems, and confer a geometric basis to recent contributions on the learning of fixed-rank non-symmetric matrices. We make connections with existing algorithms in the context of low-rank matrix completion and discuss the usefulness of the proposed framework. Numerical experiments suggest that the proposed algorithms compete with state-of-the-art algorithms and that manifold optimization offers an effective and versatile framework for the design of machine learning algorithms that learn a fixed-rank matrix. © 2013 Springer-Verlag Berlin Heidelberg.