Classification calibration dimension for general multiclass losses

We study consistency properties of surrogate loss functions for general multiclass classification problems, defined by a general loss matrix. We extend the notion of classification calibration, which has been studied for binary and multiclass 0-1 classification problems (and for certain other specific learning problems), to the general multiclass setting, and derive necessary and sufficient conditions for a surrogate loss to be classification calibrated with respect to a loss matrix in this setting. We then introduce the notion of \emph{classification calibration dimension} of a multiclass loss matrix, which measures the smallest `size' of a prediction space for which it is possible to design a convex surrogate that is classification calibrated with respect to the loss matrix. We derive both upper and lower bounds on this quantity, and use these results to analyze various loss matrices. In particular, as one application, we provide a different route from the recent result of Duchi et al.\ (2010) for analyzing the difficulty of designing `low-dimensional' convex surrogates that are consistent with respect to pairwise subset ranking losses. We anticipate the classification calibration dimension may prove to be a useful tool in the study and design of surrogate losses for general multiclass learning problems.

Cooperative Particle Swarm Optimization Based Receiver for Large-Dimension MIMO-ZPSC Systems

In this paper, we propose a cooperative particle swarm optimization (CPSO) based channel estimation/equalization scheme for multiple-input multiple-output zero-padded single-carrier (MIMO-ZPSC) systems with large dimensions in frequency selective channels. We estimate the channel state information at the receiver in time domain using a PSO based algorithm during training phase. Using the estimated channel, we perform information symbol detection in the frequency domain using FFT based processing. For this detection, we use a low complexity OLA (OverLap Add) likelihood ascent search equalizer which uses minimum mean square (MMSE) equalizer solution as the initial solution. Multiple iterations between channel estimation and data detection are carried out which significantly improves the mean square error and bit error rate performance of the receiver.

Matrix Crack Detection in Composite Plate with Spatially Random Material Properties using Fractal Dimension

Fractal dimension based damage detection method is investigated for a composite plate with random material properties. Composite material shows spatially varying random material properties because of complex manufacturing processes. Matrix cracks are considered as damage in the composite plate. Such cracks are often seen as the initial damage mechanism in composites under fatigue loading and also occur due to low velocity impact. Static deflection of the cantilevered composite plate with uniform loading is calculated using the finite element method. Damage detection is carried out based on sliding window fractal dimension operator using the static deflection. Two dimensional homogeneous Gaussian random field is generated using Karhunen-Loeve (KL) expansion to represent the spatial variation of composite material property. The robustness of fractal dimension based damage detection method is demonstrated considering the composite material properties as a two dimensional random field.

Detection and Characterization of Long-Pulse Low-Velocity Impact Damage in Plastic Bonded Explosives

Damage not only degrades the mechanical properties of explosives, but also influences the shock sensitivity, combustion and even detonation behavior of explosives. The study of impact damage is crucial in the vulnerability evaluation of explosives. A long-pulse low-velocity gas gun with a gas buffer was developed and used to induce impact damage in a hot pressed plastic bonded explosive. Various methods were used to detect and characterize the impact damage of the explosive. The microstructure was examined by use of polarized light microscopy. Fractal analysis of the micrographs was conducted by use of box counting method. The correlation between the fractal dimensions and microstructures was analyzed. Ultrasonic testing was conducted using a pulse through-transmission method to obtain the ultrasonic velocity and ultrasonic attenuation. Spectra analyses were carried out for recorded ultrasonic signals using fast Fourier transform. The correlations between the impact damage and ultrasonic parameters including ultrasonic velocities and attenuation coefficients were also analyzed. To quantitatively assess the impact induced explosive crystal fractures, particle size distribution analyses of explosive crystals were conducted by using a thorough etching technique, in which the explosives samples were soaked in a solution for enough time that the binder was totally removed. Impact induces a large extent of explosive crystal fractures and a large number of microcracks. The ultrasonic velocity decreases and attenuation coefficients increase with the presence of impact damage. Both ultrasonic parameters and fractal dimension can be used to quantitatively assess the impact damage of plastic bonded explosives.

Quantifying chaos: practical estimation of the correlation dimension

Be it a physical object or a mathematical model, a nonlinear dynamical system can display complicated aperiodic behavior, or "chaos." In many cases, this chaos is associated with motion on a strange attractor in the system's phase space. And the dimension of the strange attractor indicates the effective number of degrees of freedom in the dynamical system.

In this thesis, we investigate numerical issues involved with estimating the dimension of a strange attractor from a finite time series of measurements on the dynamical system.

Of the various definitions of dimension, we argue that the correlation dimension is the most efficiently calculable and we remark further that it is the most commonly calculated. We are concerned with the practical problems that arise in attempting to compute the correlation dimension. We deal with geometrical effects (due to the inexact self-similarity of the attractor), dynamical effects (due to the nonindependence of points generated by the dynamical system that defines the attractor), and statistical effects (due to the finite number of points that sample the attractor). We propose a modification of the standard algorithm, which eliminates a specific effect due to autocorrelation, and a new implementation of the correlation algorithm, which is computationally efficient.

Finally, we apply the algorithm to chaotic data from the Caltech tokamak and the Texas tokamak (TEXT); we conclude that plasma turbulence is not a low- dimensional phenomenon.

Identifying social determinants of urban low carbon transitions: the case of energy transition in Bilbao, Basque Country

33 p.

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.

Fabrication of a Low-Loss SSC Using High-Dose Electron Beam Lithography Exposure With Negative PMMA Resist

A silicon-on-insulator optical fiber-to-waveguide spot-size converter (SSC) using Poly-MethylMethAcrylate (PMMA) is presented for integrated optical circuits. Unlike the conventional use of PMMA as a positive resist, it has been successfully used as a negative resist with high-dose electron exposure for the fabrication of ultrafine silicon wire waveguides. Additionally, this process is able to reduce the side-wall roughness, and substantially depresses the unwanted propagation loss. Exploiting this technology, the authors demonstrated that the SSC can improve coupling efficiency by as much as over 2.5 dB per coupling facet, compared with that of SSC fabricated with PMMA as a positive resist with the same dimension.

Spectrotemporal CT data acquisition and reconstruction at low dose.

PURPOSE: X-ray computed tomography (CT) is widely used, both clinically and preclinically, for fast, high-resolution anatomic imaging; however, compelling opportunities exist to expand its use in functional imaging applications. For instance, spectral information combined with nanoparticle contrast agents enables quantification of tissue perfusion levels, while temporal information details cardiac and respiratory dynamics. The authors propose and demonstrate a projection acquisition and reconstruction strategy for 5D CT (3D+dual energy+time) which recovers spectral and temporal information without substantially increasing radiation dose or sampling time relative to anatomic imaging protocols. METHODS: The authors approach the 5D reconstruction problem within the framework of low-rank and sparse matrix decomposition. Unlike previous work on rank-sparsity constrained CT reconstruction, the authors establish an explicit rank-sparse signal model to describe the spectral and temporal dimensions. The spectral dimension is represented as a well-sampled time and energy averaged image plus regularly undersampled principal components describing the spectral contrast. The temporal dimension is represented as the same time and energy averaged reconstruction plus contiguous, spatially sparse, and irregularly sampled temporal contrast images. Using a nonlinear, image domain filtration approach, the authors refer to as rank-sparse kernel regression, the authors transfer image structure from the well-sampled time and energy averaged reconstruction to the spectral and temporal contrast images. This regularization strategy strictly constrains the reconstruction problem while approximately separating the temporal and spectral dimensions. Separability results in a highly compressed representation for the 5D data in which projections are shared between the temporal and spectral reconstruction subproblems, enabling substantial undersampling. The authors solved the 5D reconstruction problem using the split Bregman method and GPU-based implementations of backprojection, reprojection, and kernel regression. Using a preclinical mouse model, the authors apply the proposed algorithm to study myocardial injury following radiation treatment of breast cancer. RESULTS: Quantitative 5D simulations are performed using the MOBY mouse phantom. Twenty data sets (ten cardiac phases, two energies) are reconstructed with 88 μm, isotropic voxels from 450 total projections acquired over a single 360° rotation. In vivo 5D myocardial injury data sets acquired in two mice injected with gold and iodine nanoparticles are also reconstructed with 20 data sets per mouse using the same acquisition parameters (dose: ∼60 mGy). For both the simulations and the in vivo data, the reconstruction quality is sufficient to perform material decomposition into gold and iodine maps to localize the extent of myocardial injury (gold accumulation) and to measure cardiac functional metrics (vascular iodine). Their 5D CT imaging protocol represents a 95% reduction in radiation dose per cardiac phase and energy and a 40-fold decrease in projection sampling time relative to their standard imaging protocol. CONCLUSIONS: Their 5D CT data acquisition and reconstruction protocol efficiently exploits the rank-sparse nature of spectral and temporal CT data to provide high-fidelity reconstruction results without increased radiation dose or sampling time.

Urban Detection From Hyperspectral Images Using Dimension-Reduction Model and Fusion of Multiple Segmentations Based on Stuctural and Textural Features

Ce mémoire de maîtrise présente une nouvelle approche non supervisée pour détecter et segmenter les régions urbaines dans les images hyperspectrales. La méthode proposée n ́ecessite trois étapes. Tout d’abord, afin de réduire le coût calculatoire de notre algorithme, une image couleur du contenu spectral est estimée. A cette fin, une étape de réduction de dimensionalité non-linéaire, basée sur deux critères complémentaires mais contradictoires de bonne visualisation; à savoir la précision et le contraste, est réalisée pour l’affichage couleur de chaque image hyperspectrale. Ensuite, pour discriminer les régions urbaines des régions non urbaines, la seconde étape consiste à extraire quelques caractéristiques discriminantes (et complémentaires) sur cette image hyperspectrale couleur. A cette fin, nous avons extrait une série de paramètres discriminants pour décrire les caractéristiques d’une zone urbaine, principalement composée d’objets manufacturés de formes simples g ́eométriques et régulières. Nous avons utilisé des caractéristiques texturales basées sur les niveaux de gris, la magnitude du gradient ou des paramètres issus de la matrice de co-occurrence combinés avec des caractéristiques structurelles basées sur l’orientation locale du gradient de l’image et la détection locale de segments de droites. Afin de réduire encore la complexité de calcul de notre approche et éviter le problème de la ”malédiction de la dimensionnalité” quand on décide de regrouper des données de dimensions élevées, nous avons décidé de classifier individuellement, dans la dernière étape, chaque caractéristique texturale ou structurelle avec une simple procédure de K-moyennes et ensuite de combiner ces segmentations grossières, obtenues à faible coût, avec un modèle efficace de fusion de cartes de segmentations. Les expérimentations données dans ce rapport montrent que cette stratégie est efficace visuellement et se compare favorablement aux autres méthodes de détection et segmentation de zones urbaines à partir d’images hyperspectrales.

Bounding The Hochschild Cohomological Dimension

Ce mémoire a deux objectifs principaux. Premièrement de développer et interpréter les groupes de cohomologie de Hochschild de basse dimension et deuxièmement de borner la dimension cohomologique des k-algèbres par dessous; montrant que presque aucune k-algèbre commutative est quasi-libre.

A qualitative study of curriculum and instructional efforts in relation to high stakes federal and district policies in one low-performing elementary school

Les enjeux liés aux politiques éducatives ont considérablement changé au cours des dernières décennies. Ces changements sont liés, entre autres, à l’accroissement de l’imputabilité et de la reddition de compte qui est devenue une caractéristique importante des réformes curriculaires et pédagogiques. Les politiques à enjeux élevés exercent une pression énorme sur les districts et les écoles états-unienne afin qu’ils augmentent le rendement des élèves en utilisant des systèmes de conséquences (Hall & Ryan, 2011; Loeb & Strunk, 2007). Ces politiques envoient de puissants messages sur l'importance de certaines matières scolaires au détriment d'autres - circonscrivant les exigences en termes de compétences et de connaissances. La langue maternelle d’enseignement et les mathématiques sont devenues des mesures centrales sur lesquelles reposent l’évaluation et le degré de performance des districts et des écoles. Conséquemment, les administrateurs de districts et les directions d’écoles ont souvent recours à des réformes curriculaires et pédagogiques comme moyen d'augmenter le rendement des élèves dans les matières scolaires visées par ces politiques. Les politiques contraignent les acteurs scolaires de concentrer les ressources sur les programmes curriculaires et les évaluations, le développement professionnel, et la prise de décision pilotée par les données (Anagnostopoulos & Ruthledge, 2007; Honig & Hatch, 2004; Spillane, Diamond, et al., 2002; Weitz White & Rosenbaum, 2008). Cette thèse examine la manière dont les politiques à enjeux élevés opèrent quotidiennement dans les interactions et les pratiques au sein des écoles. Nous analysons plus particulièrement les différents messages provenant de la politique transmis aux acteurs scolaires sur les manières d'apporter des changements substantiels dans le curriculum et l'enseignement. Nous élargissons l’analyse en prenant en compte le rôle des administrateurs de district ainsi que des partenaires universitaires qui façonnent également la manière dont certains aspects des messages provenant des politiques sont transmis, négociés et/ou débattus et d’autres sont ignorés (Coburn & Woulfin, 2012). En utilisant l’analyse de discours, nous examinons le rôle du langage comme constituant et médiateur des interactions sociales entre les acteurs scolaires et d’autres parties prenantes. De telles analyses impliquent une investigation approfondie d’un nombre d’étude de cas limité. Les données utilisées dans cette thèse ont été colligées dans une école primaire états-unienne du mid-West. Cette étude de cas fait partie d’une étude longitudinale de quatre ans qui comprenait huit écoles dans les milieux urbains entre 1999 et 2003 (Distributed Leadership Studies, http://www.distributedleadership.org). La base de données analysée inclut des observations de réunions formelles et des entrevues auprès des administrateurs du district, des partenaires universitaires, de la direction d’école et des enseignants. En plus de l’introduction et de la problématique (chapitre 1) et de discussion et conclusion (chapitre 5), cette thèse comprend un ensemble de trois articles interdépendants. Dans le premier article (chapitre 2), nous effectuons une recension des écrits portant sur le domaine de l’implantation de politiques (policy implementation) et la complexité des relations locales, nationales et internationales dans les systèmes éducatifs. Pour démystifier cette complexité, nous portons une attention particulière à la construction de sens des acteurs scolaires comme étant une dimension clé du processus de mise en œuvre des réformes. Dans le deuxième article (chapitre 3), nous cherchons à comprendre les processus sociaux qui façonnent les réponses stratégiques des acteurs scolaires à l’égard des politiques du district et de l’état et en lien avec la mise en œuvre d’un curriculum prescrit en mathématiques. Plus particulièrement, nous explorons les différentes situations dans lesquelles les acteurs scolaires argumentent au sujet des changements curriculaires et pédagogiques proposés par les administrateurs de district et des partenaires universitaires afin d’augmenter les résultats scolaires en mathématiques dans une école à faible performance. Dans le troisième article (chapitre 4), nous cherchons à démystifier les complexités liées à l’amélioration de l’enseignement dans un environnement de politiques à enjeux élevés. Pour ce faire, nous utilisons l'interaction entre les notions d'agentivité et la structure afin d'analyser la manière dont les conceptions d’imputabilité et les idées qui découlent de l'environnement politique et les activités quotidiennes jouent dans les interactions entre les acteurs scolaires concernant sur l’enseignement de la langue maternelle. Nous explorons trois objectifs spécifiques : 1) la manière dont les politiques à enjeux élevés façonnent les éléments de l’enseignement qui sont reproduits et ceux qui sont transformés au fil du temps ; 2) la manière dont la compréhension des leaders de l’imputabilité façonne les aspects des messages politiques que les acteurs scolaires remarquent à travers les interactions et les conversations et 3) la manière les acteurs scolaires portent une attention particulière à certaines messages au détriment d’autres. Dans le dernier chapitre de cette thèse, nous discutons les forces et les limites de l’analyse secondaire de données qualitatives, les implications des résultats pour le domaine d’études de l’implantation de politiques et les pistes futures de recherches.

The social dimension: a neglected policy item in the Bologna Process

Since its beginning in 1999, the Bologna Process has influenced various aspects of higher education in its member countries, e.g., degree structures, mobility, lifelong learning, social dimension and quality assurance. The social dimension creates the focus of this research. The social dimension entered the Bologna Process agenda in 2001. Despite a decade of reforms, it somehow remained as a vague element and received low scholarly attention. This research addresses to this gap. Firstly, different meanings of the social dimension according to the major European policy actors are analysed. Unfolding the understandings of the actors revealed that the social dimension is mostly understood in terms reflecting the diversity of population on the student body accessing to, progressing in and completing higher education, with a special concern on the underrepresented groups. However, it is not possible to observe a similar commonality concerning the actual policy measures to achieve this goal. Divergence occurs with respect to the addressed underrepresented groups, i.e., all underrepresented groups or people without formal qualifications and mature learners, and the values and institutional interests traditionally promoted by these actors. Secondly, the dissertation discusses the reflection of this social dimension understanding at the national level by looking at cases of Finland, Germany and Turkey. The in-depth analyses show an awareness of the social dimension among most of the national Bologna Process actors and a common understanding of the social dimension goals. However, this understanding has not triggered action in any of the countries. The countries acted on areas which they defined problematic before the Bologna Process. Finally, based on these findings the dissertation discusses the social dimension as a policy item that managed to get into the Bologna Process agenda, but neither grew into an implementable policy, nor drop out of it. To this aim, it makes use of the multiple streams framework and explains the low agenda status social dimension with: i. the lack of a pressing problem definition: the lack of clearly defined indicators and a comprehensive monitoring system, ii. the lack of a viable solution alternative: the proposal of developing national strategies and action plans closed the way to develop generic guidelines for the social dimension to be translated into national policy processes, iii. low political perceptivity: the recent trends opt for increasing efficiency, excellence and exclusiveness discourses rather than ensuring equality and inclusiveness iv. high constraints: the social dimension by definition requires more public funding which is less appreciated and strategic constraints of the actors in allocating their resources v. the type of policy entrepreneur: the social dimension is promoted by an international stakeholder, the European Students’ Union, instead of the ministers responsible for higher education The social dimension remains a policy item in the Bologna Process which is noble enough to agree but not urgent enough to act on.

Temperamental Dimension and Anxiety Problems in a Clinical Sample of Three- to Six-year old Children: A Study of Variables

In the last few years, many researchers have studied the presence of common dimensions of temperament in subjects with symptoms of anxiety. The aim of this study is to examine the association between temperamental dimensions (high negative affect and activity level) and anxiety problems in clinical preschool children. A total of 38 children, ages 3 to 6 years, from the Infant and Adolescent Mental Health Center of Girona and the Center of Diagnosis and Early Attention of Sabadell and Olot were evaluated by parents and psychologists. Their parents completed several screening scales and, subsequently, clinical child psychopathology professionals carried out diagnostic interviews with children from the sample who presented signs of anxiety. Findings showed that children with high levels of negative affect and low activity level have pronounced symptoms of anxiety. However, children with anxiety disorders do not present different temperament styles from their peers without these pathologies

An X-ray micro-tomography system optimised for the low-dose study of living organisms

An X-ray micro-tomography system has been designed that is dedicated to the low-dose imaging of radiation sensitive living organisms and has been used to image the early development of the first few days of plant development immediately after germination. The system is based on third-generation X-ray micro-tomography system and consists of an X-ray tube, two-dimensional X-ray detector and a mechanical sample manipulation stage. The X-ray source is a 50 kVp X-ray tube with a silver target with a filter to centre the X-ray spectrum on 22 keV.A 100 mm diameter X-ray image intensifier (XRII) is used to collect the two-dimensional projection images. The rotation tomography table incorporates a linear translation mechanism to eliminate ring artefact that is commonly associated with third-generation tomography systems' Developing maize seeds (Triticum aestivum) have been imaged using the system with a cubic voxel linear dimension of 100 mum, over a diameter of 25 mm and the root lengths and volumes measured. The X-ray dose to the plants was also assessed and found to have no effect on the plant root development. (C) 2003 Elsevier Science Ltd. All rights reserved.