No description available.

Sparse approximation is a Bayesian inference program with a wide number of signal processing applications, such as Compressed Sensing recovery used in medical imaging. Previous sparse coding implementations relied on digital algorithms whose power consumption and performance scale poorly with problem size, rendering them unsuitable for portable applications, and a bottleneck in high speed applications. A novel analog architecture, implementing the Locally Competitive Algorithm (LCA), was designed and programmed onto a Field Programmable Analog Arrays (FPAAs), using floating gate transistors to set the analog parameters. A network of 6 coefficients was demonstrated to converge to similar values as a digital sparse approximation algorithm, but with better power and performance scaling. A rate encoded spiking algorithm was then developed, which was shown to converge to similar values as the LCA. A second novel architecture was designed and programmed on an FPAA implementing the spiking version of the LCA with integrate and fire neurons. A network of 18 neurons converged on similar values as a digital sparse approximation algorithm, with even better performance and power efficiency than the non-spiking network. Novel algorithms were created to increase floating gate programming speed by more than two orders of magnitude, and reduce programming error from device mismatch. A new FPAA chip was designed and tested which allowed for rapid interfacing and additional improvements in accuracy. Finally, a neuromorphic chip was designed, containing 400 integrate and fire neurons, and capable of converging on a sparse approximation solution in 10 microseconds, over 1000 times faster than the best digital solution.

For a locally compact group G, the papers [13] and [7] have many results about G-invariant subsets of G-modules, and the relationship between G-module maps, L1(G)-module maps and M(G)-module maps. In both papers, the results were given for one specific module action. In this thesis we extended many of their results to arbitrary Banach G-modules. In addition, we give detailed proofs of most of the results found in the first section of the paper [21].

For a locally compact group G, the papers [13] and [7] have many results about G-invariant subsets of G-modules, and the relationship between G-module maps, L1(G)-module maps and M(G)-module maps. In both papers, the results were given for one specific module action. In this thesis we extended many of their results to arbitrary Banach G-modules. In addition, we give detailed proofs of most of the results found in the first section of the paper [21].

Cette thèse porte sur des propriétés d'approximation généralisant la moyennabilité pour les groupes localement compacts. Ces propriétés sont définies à partir des multiplicateurs de certaines algèbres associés aux groupes. La première partie est consacrée à l'étude de la propriété p-AP, qui est une extension de la AP de Haagerup et Kraus au cadre des opérateurs sur les espaces Lp. Le résultat principal dit que les groupes de Lie simples de rang supérieur et de centre fini ne satisfont p-AP pour aucun p entre 1 et l'infini. La deuxième partie se concentre sur les multiplicateurs de Schur radiaux sur les graphes. L'étude de ces objets est motivée par les liens avec les actions de groupes discrets et la moyennabilité faible. Les trois résultats principaux donnent des conditions nécessaires et suffisantes pour qu'une fonction sur les nombres naturels définisse un multiplicateur radial sur des différentes classes de graphes généralisant les arbres. Plus précisément, les classes de graphes étudiées sont les produits d'arbres, les produits de graphes hyperboliques et les complexes cubiques CAT(0) de dimension finie. / This thesis focusses on some approximation properties which generalise amenability for locally compact groups. These properties are defined by means of multipliers of certain algebras associated to the groups. The first part is devoted to the study of the p-AP, which is an extension of the AP of Haagerup and Kraus to the context of operators on Lp spaces. The main result asserts that simple Lie groups of higher rank and finite centre do not satisfy p-AP for any p between 1 and infinity. The second part concentrates on radial Schur multipliers on graphs. The study of these objects is motivated by some connections with actions of discrete groups and weak amenability. The three main results give necessary and sufficient conditions for a function of the natural numbers to define a radial multiplier on different classes of graphs generalising trees. More precisely, the classes of graphs considered here are products of trees, products hyperbolic graphs and finite dimensional CAT(0) cube complexes.

Název práce: Geometrické lineární a nelineární problémy prostor· funkcí Autor: Petr Petráček Katedra: Katedra matematické analýzy 'kolitel: prof. RNDr. Jaroslav Lukeš, DrSc., Katedra matematické analýzy Abstrakt: Tato práce sestává ze čtyř vědeckých článk·. lánky prezentované v prvních dvou kapitolách se věnují teorii reálných a komplexních L1-preduál·. lánky prezentované v třetí a čtvrté kapitole jsou věnovány problematice line- ability a algebrability podmnožin reálných funkcí a měr. V Kapitole 1 předsta- vujeme charakterizaci komplexních L1-preduál· pomocí komplexního barycent- rického zobrazení. Tato charakterizace je přirozeným rozšířením charakterizace reálných L1 preduál· pocházející od Bednara a Laceyho. V Kapitole 2 odpoví- dáme na otázku položenou Laceym v roce 1973. Dokazujeme přitom existenci kompaktního prostoru K a uzavřeného podprostoru H ⊂ C(K) obsahujícího kon- stantní funkce, pro který platí ∂HK = K, H je maximální vzhledem k ∂HK a H není L1-preduál. V Kapitole 3 se věnujeme lineabilitě množin nikde mono- tonních znaménkových Radonových měr na Rd . Konkrétně dokazujeme existence vektorového prostoru dimenze c jehož každý nenulový prvek je nikde monotonní míra absolutně spojitá vzhledem k d-rozměrné Lebesgueově míře. Nadto dokazu- jeme, že existuje takový lineární prostor, který je hustý...

Dans cette thèse nous explorons différentes décompositions de graphes. Le titre de la présente thèse est dû au fait que la majorité de ces décompositions sont des décompositions voisin-distinguantes. En d'autres mots, nous pouvons en extraire des colorations propres des sommets. La question principale présentée dans cette thèse a été introduite par Karoński, Łuczak et Thomason: Est il possible de pondérer les arêtes d'un graphes avec les poids 1, 2 et 3, afin que tous les sommets voisins soient distingués par la somme des poids de leurs arêtes incidentes ? Cette question deviendra plus tard la fameuse 1-2-3 Conjecture. Nous présentons différentes variantes de la 1-2-3 Conjecture, ainsi que leurs liens avec les décompositions localement irrégulières. Nous nous intéressons tant à des problèmes d'optimisation qu'à des problèmes algorithmiques. Nous commençons par introduire une variante équitable des arête-pondérations voisin-somme-distinguantes, où chaque poids doit être utilisé le même nombre de fois (à l'unité près). Ensuite nous présentons une variante injective ou chaque poids est utilisé au plus une seule fois. Ce qui est un cas particulier de la variante équitable. De plus les pondérations injectives sont une variante locale des étiquetages anti-magiques. Ensuite nous modifions les conditions de distinction entre voisin en introduisant une variante 2-distinguante. les pondérations voisins-somme-2-distinguantes requierent que deux sommets voisins dans le graphe aient des sommes incidentes qui diffèrent d'au moins 2. Nous étudions le poids maximum minimal dans de telles pondérations pour certaines familles de graphes, ainsi que des problèmes de complexité. Dû aux liens entre les pondérations voisins-sommet-distinguantes et les décompositions localement irrégulières, nous nous sommes aussi intéressé à ces dernières, particulièrement pour les graphes sub-cubiques, ainsi qu'à d'autres variantes des décompositions localement irrégulières. Finalement nous présentons un jeu de pondérations à deux joueurs, ainsi qu'une théorie de décompositions qui unifie les pondérations voisin-somme-distinguantes et les décompositions localement irrégulières. / In this thesis we explore graph decompositions under different constraints. The title of the is due to the fact that most of these decompositions are neighbour-distinguishing. That is, we can extract from each such decomposition a proper vertex colouring. Moreover, most of the considered decompositions are edge partitions, and therefore can be seen as edge-colourings. The main question presented in this thesis is was introduced by Karoński, Łuczak and Thomason in [KLT04]: Can we weight the edges of a graph G, with weights 1, 2, and 3, such that any two of adjacent vertices of G are distinguished by the sum of their incident weights ? This question later becomes the famous 1-2-3 Conjecture. In this thesis we explore several variants of the 1-2-3 Conjecture, and their links with locally irregular decompositions. We are interested in both optimisation results and algorithmic problems. We first introduce an equitable version of the neighbour-sum- distinguishing edge-weightings, that is a variant where we require every edge weight to be used the same number of times up to a difference of 1. Then we explore an inject- ive variant where each edge is assigned a different weight, which yields necessarily an equitable weighting. This gives us first general upper bounds on the equitable version. Moreover, the injective variant is also a local version of the well-known antimagic la- belling. After that we explore how neighbour-sum-distinguishing weightings behave if we require sums of neighbouring vertices to differ by at least 2. Namely, we present results on the smallest maximal weight needed to construct such weightings for some classes of graphs, and study some algorithmic aspects of this problem. Due to the links between neighbour-sum-distinguishing edge weightings and locally irregular decompositions, we also explore the locally irregular index of subcubic graphs, along with other variants of the locally irregular decomposition problem. Finally, we present a more general work to- ward a general theory unifying nsd edge-weightings and locally irregular decompositions. We also present a 2-player game version of neighbour-sum-distinguishing edge-weightings and exhibit sufficient conditions for each player to win the game.

Submitted by Johnny Rodrigues (johnnyrodrigues@ufcg.edu.br) on 2018-08-08T20:21:22Z No. of bitstreams: 1 ARTHUR GILZEPH FARIAS ALMEIDA - DISSERTAÇÃO PPGMAT 2013..pdf: 508810 bytes, checksum: 02ca89b269a1cb82e4ba0a5d102acff9 (MD5) / Made available in DSpace on 2018-08-08T20:21:22Z (GMT). No. of bitstreams: 1 ARTHUR GILZEPH FARIAS ALMEIDA - DISSERTAÇÃO PPGMAT 2013..pdf: 508810 bytes, checksum: 02ca89b269a1cb82e4ba0a5d102acff9 (MD5) Previous issue date: 2013-10 / CNPq / Neste trabalho estudamos a existência de, pelo menos, três soluções distintas para dois problemas de inclusão diferencial. Para isto, faremos uso da teoria da análise convexa para funcionais localmente Lipschitz, bem como métodos variacionais. / In this work we study the existence of, at least, three distinct solutions to two problems of differential inclusion. For this, we use the theory of convex functional analysis Lipschitz locally, and variational methods.

L’objectif de cette thèse est de répondre au besoin d’accélérer la restitution des résultats de calcul d’un code CFD pour la simulation d’écoulements hydrodynamiques quasi-incompressibles. Ce code présente l’originalité de résoudre explicitement les équations de Navier-Stokes sous l’hypothèse de faible compressibilité avec des schémas numériques d’ordre élevé. Les développements effectués visent à réduire les temps de calcul à précision équivalente.Une première partie est consacrée à l’implémentation d’une formulation purement incompressible avec une résolution implicite de la pression par un schéma de projection. La formulation incompressible autorise des pas de temps plus grand en s’affranchissant de la vitesse du son, mais au prix d’une algorithmique plus complexe et de la nécessité de résoudre un système linéaire. La comparaison des deux formulations,faiblement-compressible et incompressible, tend à montrer la pertinence du schéma de projection pour les écoulements laminaires instationnaires.Un deuxième axe de développement a consisté en la proposition d’une amélioration de la méthode de frontière immergée initialement présente dans le code.Si les résultats obtenus ne sont pas encore pleinement satisfaisants, ils montrent que la montée en ordre d’une méthode de frontière immergée peut être moins contraignante en formulation incompressible.Enfin la dernière partie traite de l’immersion rapide et robuste de géométries complexes telles qu’elles peuvent être rencontrées dans l’industrie. La localisation géométrique par arbre octal permet d’évaluer rapidement une fonction de distance signée indispensable pour la méthode de frontière immergée. / An original strategy to address hydrodynamic flow was recently proposed through a high-order weakly-compressible Cartesian grid approach. The method is based on a fully-explicit temporal scheme for solving the Navier-Stokes equations. The present thesis aims to reduce the computational time required to obtain the results without deteriorating the accuracy.A first part is dedicated to the implementation of a truly incompressible formulation with an implicit solution for the pressure field through a projection scheme. The incompressible solver allows larger time step size for time integration since the speed of sound tends to infinity. In return the algorithms are no longer straight forward and a linear system has to be solved through the Pressure Poisson Equation. Comparisons carried out between both formulations show that the projection scheme can be better adapted to efficiently simulate unsteady viscous flows. Then an improvement of the immersed boundary method has been proposed. Results are not fully satisfactory for now. However, it seems easier to develop a numerical scheme for the incompressible approach rather than the weakly-compressible one.Finally, the last part addresses the setup up of complex triangulations in immersed boundary simulations. A fast and robust procedure is developed for distance computation with an octree data structure.

The main goal of this study is to propose a methodology that measures directed relations between locally stationary processes. Unlike stationary processes, locally stationary processes may present sudden pattern changes and have local characteristics in specific intervals. This behavior causes instability in measures based on Fourier transforms. The relevance of this study relies on considering these processes and propose robust methodologies that are not affected by outliers, sudden pattern changes or local behavior. We start reviewing the Partial Directed Coherence (PDC) and the Wavelet Coherence. PDC measures the directed relation between components of a multivariate stationary Vector Autoregressive (VAR) model in the frequency domain, while Wavelet Coherence is based on complex wavelets decomposition. We then propose a causal wavelet decomposition of the covariance structure for bivariate locally stationary processes: the Directed Wavelet Covariance (DWC). Compared to Fourier-based quantities, wavelet-based estimators are more appropriate for non-stationary processes and processes with local patterns, outliers and rapid regime changes like in EEG experiments with the introduction of stimuli. We then propose its estimators and calculate its expectation and analyze its variance. Next we propose a decomposition for the variance of multivariate processes with more than two components: the Partial Directed Wavelet Covariance (pDWC). Considering a N-variate locally stationary process, the pDWC calculates the Directed Wavelet Covariance of X_1(t) with X_2(t) eliminating the effect of the other components X_3(t), ... ,X_N(t). We propose two approaches to this situation. First we filter the multivariate process to remove all the exogenous influences and then we calculate the directed relation between the components. In the second case, as in Partial Directed Coherence, we consider the multivariate process as a time-varying Vector Autoregressive Model (tv-VAR) and use its coefficients in the decomposition of the covariance function to isolate the effects of the other components. We also compare results of the PDC, Wavelet Coherence and Directed Wavelet Covariance with simulated data. Finally, we present an application of the proposed Directed Wavelet Covariance and Partial Directed Wavelet Covariance on EEG data. Simulation results show that the proposed measures capture the simulated relations. The pDWC with linear filter has shown more stable estimations than the proposed pDWC considering the tv-VAR. Future studies will discuss the DWC\'s and pDWC\'s asymptotic distributions and significance tests. The proposed Directed Wavelet Covariance decomposition is a different approach to deal with non-stationary processes in the context of causality. The use of wavelets is a gain and adds to the number of studies that can be addressed when Fourier transform does not apply. The pDWC is an alternative for multivariate processes and it removes linear influences from observed external components. / O objetivo deste trabalho é propor uma metodologia para mensurar o impacto direcionado entre processos localmente estacionários. Diferente de processos estacionários, processos localmente estacionários podem apresentar mudanças bruscas e características específicas em determinados intervalos. Tal comportamento pode causar instabilidade em medidas baseadas na transformada de Fourier. A importância deste estudo se dá em englobar processos com tais características, propondo metodologias robustas que não são afetadas pela existência de mudanças bruscas, pontos discrepantes e comportamentos locais. Inicialmente apresentamos conceitos já existentes na literatura, como a Coerência Parcial Direcionada (PDC) e a Coerência de Ondaletas. A PDC mede o impacto direcionado entre componentes de um modelo vetorial autoregressivo (VAR) no domínio da frequência. A coerência de ondaletas é baseada em transformadas complexas de ondaletas. Propomos então uma decomposição no domínio de ondaletas para a estrutura de covariância de processos bivariados localmente estacionários: a Covariância Direcionada de Ondaletas (DWC). Em comparação com as quantidades baseadas na tranformada Fourier, os estimadores baseados em ondaletas são mais apropriados para processos não estacionários com padrões locais, pontos discrepantes ou mudanças rápidas de regime, como em experimentos de eletroencefalograma (EEG) com a introdução de estímulo. Ainda, propomos um estimador para a DWC, calculamos a esperança deste estimador e avaliamos sua variância. Em seguida, propomos uma quantidade análoga à DWC para processos multivariados com mais de duas componentes: a Covariância Parcial Direcionada de Ondaletas (pDWC). Considerando um processo N-variado localmente estacionário, a pDWC calcula a Covariância Direcionada de Ondaletas entre X_1(t) e X_2(t) eliminando o efeito das outras componentes X_3(t), ... , X_N(t). Propomos duas abordagens para a pDWC: na primeira, a pDWC é calculada após a aplicação de um filtro linear que remove o efeito das variáveis exógenas. No segundo caso, a exemplo da Coerência Parcial Direcionada, consideramos o processo multivariado como um Modelo Autoregressivo de Vetorial variante no tempo (tv-VAR) e usamos seus coeficientes na decomposição da função de covariância para isolar os efeitos das demais componentes. Também comparamos os resultados da PDC, Coerência de Ondaletas e Covariância Direcionada de Ondaletas com dados simulados. Por fim, apresentamos uma aplicação da DWC e da pDWC em dados de EEG. Identificamos nas simulações que tanto as medidas já existentes na literatura quanto as quantidades propostas identificaram as relações simuladas. A pDWC proposta com filtros lineares apresentou estimações mais estáveis do que a pDWC considerando os modelos tv-VAR. Estudos futuros discutirão as propriedades assintóticas e testes de significância da DWC e pDWC.