Regression on fixed-rank positive semidefinite matrices: A Riemannian approach

The paper addresses the problem of learning a regression model parameterized by a fixed-rank positive semidefinite matrix. The focus is on the nonlinear nature of the search space and on scalability to high-dimensional problems. The mathematical developments rely on the theory of gradient descent algorithms adapted to the Riemannian geometry that underlies the set of fixedrank positive semidefinite matrices. In contrast with previous contributions in the literature, no restrictions are imposed on the range space of the learned matrix. The resulting algorithms maintain a linear complexity in the problem size and enjoy important invariance properties. We apply the proposed algorithms to the problem of learning a distance function parameterized by a positive semidefinite matrix. Good performance is observed on classical benchmarks. © 2011 Gilles Meyer, Silvere Bonnabel and Rodolphe Sepulchre.

Generalized power method for sparse principal component analysis

In this paper we develop a new approach to sparse principal component analysis (sparse PCA). We propose two single-unit and two block optimization formulations of the sparse PCA problem, aimed at extracting a single sparse dominant principal component of a data matrix, or more components at once, respectively. While the initial formulations involve nonconvex functions, and are therefore computationally intractable, we rewrite them into the form of an optimization program involving maximization of a convex function on a compact set. The dimension of the search space is decreased enormously if the data matrix has many more columns (variables) than rows. We then propose and analyze a simple gradient method suited for the task. It appears that our algorithm has best convergence properties in the case when either the objective function or the feasible set are strongly convex, which is the case with our single-unit formulations and can be enforced in the block case. Finally, we demonstrate numerically on a set of random and gene expression test problems that our approach outperforms existing algorithms both in quality of the obtained solution and in computational speed. © 2010 Michel Journée, Yurii Nesterov, Peter Richtárik and Rodolphe Sepulchre.

Consensus on homogeneous manifolds

The present paper considers distributed consensus algorithms for agents evolving on a connected compact homogeneous (CCH) manifold. The agents track no external reference and communicate their relative state according to an interconnection graph. The paper first formalizes the consensus problem for synchronization (i.e. maximizing the consensus) and balancing (i.e. minimizing the consensus); it thereby introduces the induced arithmetic mean, an easily computable mean position on CCH manifolds. Then it proposes and analyzes various consensus algorithms on manifolds: natural gradient algorithms which reach local consensus equilibria; an adaptation using auxiliary variables for almost-global synchronization or balancing; and a stochastic gossip setting for global synchronization. It closes by investigating the dependence of synchronization properties on the attraction function between interacting agents on the circle. The theory is also illustrated on SO(n) and on the Grassmann manifolds. ©2009 IEEE.

Synchronization with partial state feedback on SO(n)

In this paper we consider the problem of constructing a distributed feedback law to achieve synchronization for a group of k agents whose states evolve on SO(n) and which exchange only partial state information along communication links. The partial state information is given by the action of the state on reference vectors in ℝn. We propose a gradient based control law which achieves exponential local convergence to a synchronization configuration under a rank condition on a generalized Laplacian matrix. Furthermore, we discuss the case of time-varying reference vectors and provide a convergence result for this case. The latter helps reach synchronization, requiring less communication links and weaker conditions on the instantaneous reference vectors. Our methods are illustrated on an attitude synchronization problem where agents exchange only their relative positions observed in the respective body frames. ©2009 IEEE.

Consensus optimization on manifolds

The present paper considers distributed consensus algorithms that involve N agents evolving on a connected compact homogeneous manifold. The agents track no external reference and communicate their relative state according to a communication graph. The consensus problem is formulated in terms of the extrema of a cost function. This leads to efficient gradient algorithms to synchronize (i.e., maximizing the consensus) or balance (i.e., minimizing the consensus) the agents; a convenient adaptation of the gradient algorithms is used when the communication graph is directed and time-varying. The cost function is linked to a specific centroid definition on manifolds, introduced here as the induced arithmetic mean, that is easily computable in closed form and may be of independent interest for a number of manifolds. The special orthogonal group SO (n) and the Grassmann manifold Grass (p, n) are treated as original examples. A link is also drawn with the many existing results on the circle. © 2009 Society for Industrial and Applied Mathematics.

Continuous dynamical systems that realize discrete optimization on the hypercube

We study the problem of finding a local minimum of a multilinear function E over the discrete set {0,1}n. The search is achieved by a gradient-like system in [0,1]n with cost function E. Under mild restrictions on the metric, the stable attractors of the gradient-like system are shown to produce solutions of the problem, even when they are not in the vicinity of the discrete set {0,1}n. Moreover, the gradient-like system connects with interior point methods for linear programming and with the analog neural network studied by Vidyasagar (IEEE Trans. Automat. Control 40 (8) (1995) 1359), in the same context. © 2004 Elsevier B.V. All rights reserved.

The inverse problem in magnetic force microscopy--inferring sample magnetization from MFM images.

Nanomagnetic structures have the potential to surpass silicon's scaling limitations both as elements in hybrid CMOS logic and as novel computational elements. Magnetic force microscopy (MFM) offers a convenient characterization technique for use in the design of such nanomagnetic structures. MFM measures the magnetic field and not the sample's magnetization. As such the question of the uniqueness of the relationship between an external magnetic field and a magnetization distribution is a relevant one. To study this problem we present a simple algorithm which searches for magnetization distributions consistent with an external magnetic field and solutions to the micromagnetic equations' qualitative features. The algorithm is not computationally intensive and is found to be effective for our test cases. On the basis of our results we propose a systematic approach for interpreting MFM measurements.

Obtaining bifurcation diagrams with a thermoacoustic network model

Linear techniques can predict whether the non-oscillating (steady) state of a thermoacoustic system is stable or unstable. With a sufficiently large impulse, however, a thermoacoustic system can reach a stable oscillating state even when the steady state is also stable. A nonlinear analysis is required to predict the existence of this oscillating state. Continuation methods are often used for this but they are computationally expensive. In this paper, an acoustic network code called LOTAN is used to obtain the steady and the oscillating solutions for a horizontal Rijke tube. The heat release is modelled as a nonlinear function of the mass flow rate. Several test cases from the literature are analysed in order to investigate the effect of various nonlinear terms in the flame model. The results agree well with the literature, showing that LOTAN can be used to map the steady and oscillating solutions as a function of the control parameters. Furthermore, the nature of the bifurcation between steady and oscillating states can be predicted directly from the nonlinear terms inside the flame model. Copyright © 2012 by ASME.

Application of the multi-objective Alliance algorithm to a benchmark aerodynamic optimization problem

This paper introduces a new version of the multiobjective Alliance Algorithm (MOAA) applied to the optimization of the NACA 0012 airfoil section, for minimization of drag and maximization of lift coefficients, based on eight section shape parameters. Two software packages are used: XFoil which evaluates each new candidate airfoil section in terms of its aerodynamic efficiency, and a Free-Form Deformation tool to manage the section geometry modifications. Two versions of the problem are formulated with different design variable bounds. The performance of this approach is compared, using two indicators and a statistical test, with that obtained using NSGA-II and multi-objective Tabu Search (MOTS) to guide the optimization. The results show that the MOAA outperforms MOTS and obtains comparable results with NSGA-II on the first problem, while in the other case NSGA-II is not able to find feasible solutions and the MOAA is able to outperform MOTS. © 2013 IEEE.

Finite Element studies on indentation size effect using a higher order strain gradient theory

In this work, a Finite Element implementation of a higher order strain gradient theory (due to Fleck and Hutchinson, 2001) has been used within the framework of large deformation elasto-viscoplasticity to study the indentation of metals with indenters of various geometries. Of particular interest is the indentation size effect (ISE) commonly observed in experiments where the hardness of a range of materials is found to be significantly higher at small depths of indentation but reduce to a lower, constant value at larger depths. That the ISE can be explained by strain gradient plasticity is well known but this work aims to qualitatively compare a gamut of experimental observations on this effect with predictions from a higher order strain gradient theory. Results indicate that many of the experimental observations are qualitatively borne out by our simulations. However, areas exist where conflicting experimental results make assessment of numerical predictions difficult. © 2012 Elsevier Ltd. All rights reserved.

Optimal data fitting on lie groups: A coset approach

This work considers the problem of fitting data on a Lie group by a coset of a compact subgroup. This problem can be seen as an extension of the problem of fitting affine subspaces in n to data which can be solved using principal component analysis. We show how the fitting problem can be reduced for biinvariant distances to a generalized mean calculation on an homogeneous space. For biinvariant Riemannian distances we provide an algorithm based on the Karcher mean gradient algorithm. We illustrate our approach by some examples on SO(n). © 2010 Springer -Verlag Berlin Heidelberg.

Scalar gradient behaviour in MILD combustion

The results of three-dimensional Direct Numerical Simulation (DNS) of Moderate, Intense Low-oxygen Dilution (MILD) and conventional premixed turbulent combustion conducted using a skeletal mechanism including the effects of non-unity Lewis numbers and temperature dependent transport properties are analysed to investigate combustion characteristics using scalar gradient information. The DNS data is also used to synthesise laser induced fluorescence (LIF) signals of OH, CH2O, and CHO. These signals are analysed to verify if they can be used to study turbulent MILD combustion and it has been observed that at least two (OH and CH2O) LIF signals are required since the OH increase across the reaction zone is smaller in MILD combustion compared to premixed combustion. The scalar gradient PDFs conditioned on the reaction rate obtained from the DNS data and synthesised LIF signals suggests a strong gradient in the direction normal to the MILD reaction zone with moderate reaction rate implying flamelet combustion. However, the PDF of the normal gradient is as broad as for the tangential gradient when the reaction rate is high. This suggests a non-flamelet behaviour, which is due to interaction of reaction zones. The analysis of the conditional PDFs for the premixed case confirms the expected behaviour of scalar gradient in flamelet combustion. It has been shown that the LIF signals synthesised using 2D slices of DNS data also provide very similar insights. These results demonstrate that the so-called flameless combustion is not an idealised homogeneous reactive mixture but has common features of conventional combustion while containing distinctive characteristics. © 2013 The Combustion Institute.