The structural vibration damping behaviour of plates and beams can be improved by the application of viscoelastic passive damping materials. Unconstrained layer damping treatments applied to metal plate systems were studied experimentally. Design and modelling of novel fibre reinforced constrained layer damping materials was performed, and implementation of these composite damping materials into laminated composite sandwich constructions commonly used as structural elements within large composite marine vessels was explored. These studies established effective methods for examining, designing and applying damping materials to metal and composite marine structures. Two test fixtures were designed and constructed to facilitate testing of viscoelastic material damping properties to ISO 6721-3 and ASTM E756. Values of material damping made in accordance with ASTM E756 over a range of temperatures were compared to values produced by a Dynamic Mechanical Analyser (DMA). Glass transition temperatures and peak damping values were found to agree well, although results deviated significantly at temperatures above the glass transition temperature. The relative influence of damping layer thickness, ambient temperature, edge conditions, plate dimensions and substrate material on the system damping performance of metal plates treated with an unconstrained viscoelastic layer was investigated experimentally. This investigation found that substrate material had the greatest influence on system damping performance, followed by damping layer thickness and plate size. Plate edge conditions were found to have little influence on the measured system damping performance. These results were dependent on the values of each variable used in the study. Modal damping behaviour of a novel fibre reinforced composite constrained layer damping material was investigated using finite element analysis and experimental methods. The material consisted of two carbon fibre reinforced polymer (CFRP) layers surrounding a viscoelastic core. Opposing complex sinusoidal fibre patterns in the CFRP face sheets were used to achieve stress-coupling by way of orthotropic anisotopy about the core. A finite element model was developed in MATLAB to determine the modal damping, displacement, stress, and strain behaviour of these complex patterned fibre constrained layer damping (CPF-CLD) materials. This model was validated using experimental results produced by modal damping measurements on CPF-CLD beam test specimens. Studies of multiple fibre pattern arrangements found that fibre pattern properties and the resulting localised material property distributions influenced modal damping performance. Inclusion of CPF-CLD materials in laminated composite sandwich geometries commonly used in marine hull and bulkhead constructions was studied experimentally. Composite sandwich beam test specimens were fabricated using materials and techniques frequently used in industry. It was found that the greatest increases in modal damping performance were achieved when the CPF-CLD materials were applied to bulkhead geometries, and were inserted within the sandwich structure, rather than being attached to the surface.

A number of researchers have shown that ear recognition is a viable alternative to more common biometrics such as fingerprint, face and iris because the ear is relatively stable over time, the ear is non-invasive to capture, the ear is expressionless, and both the ear’s geometry and shape have significant variation among individuals. Researchers have used different approaches to enhance ear recognition. Some researchers have improved upon existing algorithms, some have developed algorithms from scratch to assist with recognizing individuals by ears, and some researchers have taken algorithms tried and tested for another purpose, i.e., face recognition, and applied them to ear recognition. These approaches have resulted in a number of state-of-the-art effective methods to identify individuals by ears. However, most ear recognition research has been done using ear images that were captured in an ideal setting: ear images have near perfect lighting for image quality, ears are in the same position for each subject, and ears are without earrings, hair occlusions, or anything else that could block viewing of the entire ear. In order for ear recognition to be practical, current approaches must be improved. Ear recognition must move beyond ideal settings and demonstrate effectiveness in an unconstrained environment reflective of real world conditions. Ear recognition approaches must be scalable to handle large groups of people. And, ear recognition should demonstrate effectiveness across a diverse population. This dissertation advances ear recognition from ideal settings to real world settings. We devised an ear recognition framework that outperformed state-of-the-art recognition approaches using the most challenging sets of publicly available ear images and the most voluminous set of unconstrained ear images that we are aware of. We developed a Convolutional Neural Network-based solution for ear normalization and description, we designed a two-stage landmark detector, and we fused learned and handcrafted descriptors. Using our framework, we identified some individuals that are wearing earrings and that have other occlusions, such as hair. The results suggest that our framework can be a gateway for identification of individuals in real world conditions.

[pt] A Programação Quadrática Binária Irrestrita (UBQP) é amplamente estudada. Trata-se de uma ferramenta de modelagem poderosa, mas otimizar de um problema NP-difícil. Neste trabalho uma nova abordagem é apresentada, que pode ser usada para construir um algoritmo exato. Além disso, a ideia básica que fundamenta o trabalho pode ser usado em um espectro ainda mais amplo de problemas. O algoritmo exato derivado do novo método é altamente paralelizável, o que é uma característica desejável nos dias de hoje em que cloud computing já é uma realidade. Para instâncias razoavelmente grandes do UBQP, o novo método pode paralelizar a centenas, ou até milhares, de núcleos com facilidade, com um aumento de desempenho quase linear. / [en] Unconstrained Binary Quadratic Programming (UBQP) is widely studied. It is a powerful modeling tool and its associate problem is NP-hard. In this work a new approach is introduced, which can be used to build an exact algorithm. Also, the fundamental idea behind it can be used in an even wider family of problems. This exact algorithm derived from the new method is highly parallelizable, which is a desired feature nowadays, when the cloud computing is a reality. For reasonably large instances of UBQP, the new method can parallelize to hundreds, or even thousands, of cores easily, with a near-linear speedup.

Companies have long been interested in gauging the customer’s level of interest in their advertisements. By analyzing the gaze direction of individuals viewing a public advertisement, we can infer their level of engagement. Head pose detection allows us to deduce pertinent information about gaze direction. Using video sensors, machine learning methods, and image processing techniques, information pertaining to the head pose of people viewing advertisements can be automatically collected and mined. We propose a method for the coarse classification of head pose from low-resolution images in crowded scenes captured through a single camera and under different lighting conditions. Our method improves on the technique described in [1]; we introduce several modifications to the latter scheme to improve classification accuracy. First, we devise a mechanism that uses a cascade of three binary Support Vector Machines (SVM) classifiers instead of a single multi-class classifier. Second, we employ a bigger dataset for training by combining eight publicly available databases. Third, we use two sets of appearance features, Similarity Distance Map (SDM) and Gabor Wavelet (GW), to train the SVM classifiers. The scheme is tested with cross validation using the dataset and on videos we collected in a lab experiment. We found a significant improvement in the results achieved by the proposed method over existing schemes, especially for video pose classification. The results show that the proposed method is more robust under varying light conditions and facial expressions and in the presences of facial accessories compared to [1].

Gaze tracking has already shown to be a popular technology for desktop devices. When it comes to gaze tracking for mobile devices, however, there is still a lot of progress to be made. There’s still no high accuracy gaze tracking available that works in an unconstrained setting for mobile devices. This work makes contributions in the area of appearance-based unconstrained gaze estimation. Artificial neural networks are trained on GazeCapture, a publicly available dataset for mobile gaze estimation containing over 2 million face images and corresponding gaze labels. In this work, Siamese neural networks are trained to learn linear distances between face images for different gaze points. Then, during inference, calibration points are used to estimate gaze points. This approach is shown to be an effective way of utilizing calibration points in order to improve the result of gaze estimation. / Ögonblickspårning har redan etablerat sig som en populär teknologi för stationära enheter. När det dock gäller mobila enheter så finns det framsteg att göra. Det saknas fortfarande en lösning för ögonblickspårning som fungerar i en undantagsfri miljö för mobila enheter. Detta examensarbete ämnar att bidra till en sådan lösning. Artificiella neurala nätverk tränas på GazeCapture, en allmänt tillgänglig datasamling som består av över 2 miljoner ansiktsbilder samt korresponderande etikett för ögonblickspunkt. I detta examensarbete tränas Siamesiska neurala nätverk för att lära sig det linjära avståndet mellan två ögonblickspunkter. Sedan utnyttjas en samling med kalibreringsbilder för att estimera ögonblickspunkter. Denna teknik visar sig vara ett effektivt sätt att nyttja kalibreringsbilder med målet att förbättra resultatet för ögonblicksestimering.

A main goal of mathematics educators is to guide students in becoming better problem solvers; however, the recipe for successful problem solving is complex due to the varying factors that play a role in the problem solving process (Schoenfeld, 1992). There is a limited amount of research that examines problem solving when students work on non-routine problems outside of the classroom; therefore, the goal of this study is to use secondary data analysis to discover what factors (Schoenfeld, 1992) relate to problem solving on non-routine unconstrained tasks of students in the middle grades. Identifying the factors that relate to successful unconstrained non-routine problem solving can help mathematics teachers and policy makers make more informed decisions about curriculum and instruction in order to enhance problem solving aptitude. Using Schoenfeld’s (1992) theoretical framework for mathematical behavior, the following question set the groundwork for the current study: What resource (computational skills and heuristics), control (self-regulation), and belief/affect factors (demographics, motivation, and anxiety) both individually and collectively relate to unconstrained non-routine mathematical problem solving? The research question is answered in a series of three stages that examines how the factors relate to a) problem correctness, b) correct problem set-up, and c) problem completion. Results suggest that higher levels of self-regulation, and SES status predict problem completion; higher self-regulation, ability beliefs, and SES predict correctly setting-up the problem; and higher levels of anxiety and stronger computational skills predict solving the problem correctly. Reasons for the patterns of results are discussed, as well as suggestions for future research to extend on the current findings. / Math & Science Education

The research is a study of the efficiency and robustness of genetic algorithm to instances of both discrete and continuous global optimization problems. We developed genetic algorithm based heuristics to find the global minimum to problem instances considered. In the discrete category, we considered two instances of real-world space allocation problems that arose from an academic environment in a developing country. These are the university timetabling problem and hostel space allocation problem. University timetabling represents a difficult optimization problem and finding a high quality solution is a challenging task. Many approaches, based on instances from developed countries, have been reported in the literature. However, most developing countries are yet to appreciate the deployment of heuristics and metaheuristics in handling the timetabling problem. We therefore worked on an instance from a university in Nigeria to show the feasibility and efficiency of heuristic method to the timetabling problem. We adopt a simplified bottom up approach in which timetable are build around departments. Thus a small portion of real data was used for experimental testing purposes. As with similar baseline studies in literature, we employ genetic algorithm to solve this instance and show that efficient solutions that meet stated constraints can be obtained with the metaheuristics. This thesis further focuses on an instance of university space allocation problem, namely the hostel space allocation problem. This is a new instance of the space allocation problems that has not been studied by metaheuristic researchers to the best of our knowledge. The problem aims at the allocation of categories of students into available hostel space. This must be done without violating any hard constraints but satisfying as many soft constraints as possible and ensuring optimum space utilization. We identified some issues in the problem that helped to adapt metaheuristic approach to solve it. The problem is multi-stage and highly constrained. We first highlight an initial investigation based on genetic algorithm adapted to find a good solution within the search space of the hostel space allocation problem. Some ideas are introduced to increase the overall performance of initial results based on instance of the problem from our case study. Computational results obtained are reported to demonstrate the effectiveness of the solution approaches employed. Sensitivity analysis was conducted on the genetic algorithm for the two SAPs considered to determine the best parameter values that consistently give good solutions. We noted that the genetic algorithms perform well specially, when repair strategies are incorporated. This thesis pioneers the application of metaheuristics to solve the hostel space allocation problem. It provides a baseline study of the problem based on genetic algorithms with associated test data sets. We report the best known results for the test instances. It is a known fact that many real-life problems are formulated as global optimization problems with continuous variables. On the continuous global optimization category therefore, we focus on improving the efficiency and reliability of real coded genetic algorithm for solving unconstrained global optimization, mainly through hybridization with exploratory features. Hybridization has widely been recognized as one of the most attractive approach to solving unconstrained global optimization. Literatures have shown that hybridization helps component heuristics to taking advantage of their individual strengths while avoiding their weaknesses. We therefore derived three modified forms of real coded genetic algorithm by hybridizing the standard real-coded genetic algorithm with pattern search and vector projection. These are combined to form three new algorithms namely, RCGA-PS, RCGA-P, and RCGA-PS-P. The hybridization strategy used and results obtained are reported and compared with the standard real-coded genetic algorithm. Experimental studies show that all the modified algorithms perform better than the original algorithm.

Η συνεισφορά της παρούσας διατριβής επικεντρώνεται στην ανάπτυξη και στη Μαθηματική θεμελίωση νέων μεθόδων συζυγών κλίσεων για βελτιστοποίηση χωρίς περιορισμούς και στη μελέτη νέων μεθόδων εκπαίδευσης νευρωνικών δικτύων και εφαρμογών τους. Αναπτύσσουμε δύο νέες μεθόδους βελτιστοποίησης, οι οποίες ανήκουν στην κλάση των μεθόδων συζυγών κλίσεων. Οι νέες μέθοδοι βασίζονται σε νέες εξισώσεις της τέμνουσας με ισχυρά θεωρητικά πλεονεκτήματα, όπως η προσέγγιση με μεγαλύτερη ακρίβεια της επιφάνεια της αντικειμενικής συνάρτησης. Επιπλέον, μία σημαντική ιδιότητα και των δύο προτεινόμενων μεθόδων είναι ότι εγγυώνται επαρκή μείωση ανεξάρτητα από την ακρίβεια της γραμμικής αναζήτησης, αποφεύγοντας τις συχνά αναποτελεσματικές επανεκκινήσεις. Επίσης, αποδείξαμε την ολική σύγκλιση των προτεινόμενων μεθόδων για μη κυρτές συναρτήσεις. Με βάση τα αριθμητικά μας αποτελέσματα καταλήγουμε στο συμπέρασμα ότι οι νέες μέθοδοι έχουν πολύ καλή υπολογιστική αποτελεσματικότητα, όπως και καλή ταχύτητα επίλυσης των προβλημάτων, υπερτερώντας σημαντικά των κλασικών μεθόδων συζυγών κλίσεων. Το δεύτερο μέρος της διατριβής είναι αφιερωμένο στην ανάπτυξη και στη μελέτη νέων μεθόδων εκπαίδευσης νευρωνικών δικτύων. Προτείνουμε νέες μεθόδους, οι οποίες διατηρούν τα πλεονεκτήματα των κλασικών μεθόδων συζυγών κλίσεων και εξασφαλίζουν τη δημιουργία κατευθύνσεων μείωσης αποφεύγοντας τις συχνά αναποτελεσματικές επανεκκινήσεις. Επιπλέον, αποδείξαμε ότι οι προτεινόμενες μέθοδοι συγκλίνουν ολικά για μη κυρτές συναρτήσεις. Τα αριθμητικά αποτελέσματα επαληθεύουν ότι οι προτεινόμενες μέθοδοι παρέχουν γρήγορη, σταθερότερη και πιο αξιόπιστη σύγκλιση, υπερτερώντας των κλασικών μεθόδων εκπαίδευσης. Η παρουσίαση του ερευνητικού μέρους της διατριβής ολοκληρώνεται με μία νέα μέθοδο εκπαίδευσης νευρωνικών δικτύων, η οποία βασίζεται σε μία καμπυλόγραμμη αναζήτηση. Η μέθοδος χρησιμοποιεί τη BFGS ενημέρωση ελάχιστης μνήμης για τον υπολογισμό των κατευθύνσεων μείωσης, η οποία αντλεί πληροφορία από την ιδιοσύνθεση του προσεγγιστικού Eσσιανού πίνακα, αποφεύγοντας οποιαδήποτε αποθήκευση ή παραγοντοποίηση πίνακα, έτσι ώστε η μέθοδος να μπορεί να εφαρμοστεί για την εκπαίδευση νευρωνικών δικτύων μεγάλης κλίμακας. Ο αλγόριθμος εφαρμόζεται σε προβλήματα από το πεδίο της τεχνητής νοημοσύνης και της βιοπληροφορικής καταγράφοντας πολύ καλά αποτελέσματα. Επίσης, με σκοπό την αύξηση της ικανότητας γενίκευσης των εκπαιδευόμενων δικτύων διερευνήσαμε πειραματικά και αξιολογήσαμε την εφαρμογή τεχνικών μείωσης της διάστασης δεδομένων στην απόδοση της γενίκευσης των τεχνητών νευρωνικών δικτύων σε μεγάλης κλίμακας δεδομένα βιοϊατρικής. / The contribution of this thesis focuses on the development and the Mathematical foundation of new conjugate gradient methods for unconstrained optimization and on the study of new neural network training methods and their applications. We propose two new conjugate gradient methods for unconstrained optimization. The proposed methods are based on new secant equations with strong theoretical advantages i.e. they approximate the surface of the objective function with higher accuracy. Moreover, they have the attractive property of ensuring sufficient descent independent of the accuracy of the line search, avoiding thereby the usual inefficient restarts. Further, we have established the global convergence of the proposed methods for general functions under mild conditions. Based on our numerical results we conclude that our proposed methods outperform classical conjugate gradient methods in both efficiency and robustness. The second part of the thesis is devoted on the study and development of new neural network training algorithms. More specifically, we propose some new training methods which preserve the advantages of classical conjugate gradient methods while simultaneously ensure sufficient descent using any line search, avoiding thereby the usual inefficient restarts. Moreover, we have established the global convergence of our proposed methods for general functions. Encouraging numerical experiments on famous benchmarks verify that the presented methods provide fast, stable and reliable convergence, outperforming classical training methods. Finally, the presentation of the research work of this dissertation is fulfilled with the presentation of a new curvilinear algorithm for training large neural networks which is based on the analysis of the eigenstructure of the memoryless BFGS matrices. The proposed method preserves the strong convergence properties provided by the quasi-Newton direction while simultaneously it exploits the nonconvexity of the error surface through the computation of the negative curvature direction without using any storage and matrix factorization. Our numerical experiments have shown that the proposed method outperforms other popular training methods on famous benchmarks. Furthermore, for improving the generalization capability of trained ANNs, we explore the incorporation of several dimensionality reduction techniques as a pre-processing step. To this end, we have experimentally evaluated the application of dimensional reduction techniques for increasing the generalization capability of neural network in large biomedical datasets.

In this thesis, we investigate methods of finding a local minimum for unconstrained problems of non-convex functions with n variables, by following the solution curve of a system of ordinary differential equations. The motivation for this was the fact that existing methods (e.g. those based on Newton methods with line search) sometimes terminate at a non-stationary point when applied to functions f(x) that do not a have positive-definite Hessian (i.e. ∇²f → 0) for all x. Even when methods terminate at a stationary point it could be a saddle or maximum rather than a minimum. The only method which makes intuitive sense in non-convex region is the trust region approach where we seek a step which minimises a quadratic model subject to a restriction on the two-norm of the step size. This gives a well-defined search direction but at the expense of a costly evaluation. The algorithms derived in this thesis are gradient based methods which require systems of equations to be solved at each step but which do not use a line search in the usual sense. Progress along the Continuous Steepest Descent Path (CSDP) is governed both by the decrease in the function value and measures of accuracy of a local quadratic model. Numerical results on specially constructed test problems and a number of standard test problems from CUTEr [38] show that the approaches we have considered are more promising when compared with routines in the optimization tool box of MATLAB [46], namely the trust region method and the quasi-Newton method. In particular, they perform well in comparison with the, superficially similar, gradient-flow method proposed by Behrman [7].

Many attempts have been made to determine what factors are most important in one’s potential for learning to read. These have been investigated for the purpose of determining effective instructional procedures. This chapter on literacy development aims to problematize instructional approaches that emphasize children’s deficits and, instead, make more visible approaches that build on children’s strengths, regardless of their individual differences. A fundamental belief must be that literacy is teachable and, for all children, learnable. This chapter will delineate those aspects of literacy that are malleable, evidenced by case studies and research review, and will suggest pathways (i.e., skills, strategies, and procedures) that are most effective for teachers and researchers to pursue. It describes connections to the broad field of deafness and literacy development and implications for teachers, administrators, service providers, and others concerned with the literacy success of deaf and hard-of-hearing children.