WICKRAMARATHNE, T. L. (M.S., Electrical and Computer Engineering) A Belief Theoretic Approach for Automated Collaborative Filtering (May 2008) Abstract of a thesis at the University of Miami. Thesis supervised by Professor Kamal Premaratne. No. of pages in text. (84) Automated Collaborative Filtering (ACF) is one of the most successful strategies available for recommender systems. Application of ACF in more sensitive and critical applications however has been hampered by the absence of better mechanisms to accommodate imperfections (ambiguities and uncertainties in ratings, missing ratings, etc.) that are inherent in user preference ratings and propagate such imperfections throughout the decision making process. Thus one is compelled to make various "assumptions" regarding the user preferences giving rise to predictions that lack sufficient integrity. With its Dempster-Shafer belief theoretic basis, CoFiDS, the automated Collaborative Filtering algorithm proposed in this thesis, can (a) represent a wide variety of data imperfections; (b) propagate the partial knowledge that such data imperfections generate throughout the decision-making process; and (c) conveniently incorporate contextual information from multiple sources. The "soft" predictions that CoFiDS generates provide substantial exibility to the domain expert. Depending on the associated DS theoretic belief-plausibility measures, the domain expert can either render a "hard" decision or narrow down the possible set of predictions to as smaller set as necessary. With its capability to accommodate data imperfections, CoFiDS widens the applicability of ACF, from the more popular domains, such as movie and book recommendations, to more sensitive and critical problem domains, such as medical expert support systems, homeland security and surveillance, etc. We use a benchmark movie dataset and a synthetic dataset to validate CoFiDS and compare it to several existing ACF systems.

No description available.

No description available.

In this study, a numerical model is developed to determine shoreline changes due to wind wave induced longshore sediment transport, by solving sediment continuity equation and taking one line theory as a base, in existence of seawalls, groins, T-groins, offshore breakwaters and beach nourishment projects, whose dimensions and locations may be given arbitrarily. The model computes the transformation of deep water wave characteristics up to the surf zone and eventually gives the result of shoreline changes with user-friendly visual outputs. A method of representative wave input as annual average wave characteristics is presented. Compatibility of the currently developed tool is tested by a case study and it is shown that the results, obtained from the model, are in good agreement qualitatively with field measurements. In the scope of this study, input manner of long term annual wave data into model in miscellaneous ways is also discussed.

This paper presents a new approach for characterizing heat sealing processes in packaging machines, which identifies all appropriate and optimum sealing parameters subject to given quality criteria. The presented approach is distinct from established methods, such as heat seal curves or response surface plots, in three aspects: It provides an easy-to-read chart indicating optimum and appropriate settings of sealing temperature and dwell time, as well as process robustness for each parameter set. The data points of the charts are generated directly from experiments without using any regression model. The design of experiments features a determinate sequence of tests and no iterative loops are necessary. The approach stems from a ‘trade-off’ model of heat sealing and a specific multi-objective optimization method. This trade-off model states that improving seam quality, reducing sealing temperature, reducing dwell time, and increasing process robustness are conflicting objectives. This hypothesis is deduced from a literature review and verified by experiments. The optimization method claims that all optimum sealing parameters are part of the solution of a multi-objective optimization problem, the Pareto-frontier. This hypothesis is elaborated into a design of experiments using the normal-boundary-intersection method, and is verified using a peelable seam and tear seam scenario.

The subject of this thesis is the generation of spatial temporal structures in living cells. Specifically, we studied the Min-system in the bacterium Escherichia coli. It consists of the MinC, the MinD, and the MinE proteins, which play an important role in the correct selection of the cell division site. The Min-proteins oscillate between the two cell poles and thereby prevent division at these locations. In this way, E. coli divides at the center, producing two daughter cells of equal size, providing them with the complete genetic patrimony. Our goal is to perform a quantitative study, both theoretical and experimental, in order to reveal the mechanism underlying the Min-oscillations. Experimentally, we characterize theMin-system, measuring the temporal period of the oscillations as a function of the cell length, the time-averaged protein distributions, and the in vivo Min-protein mobility by means of different fluorescence microscopy techniques. Theoretically, we discuss a deterministic description based on the exchange of Minproteins between the cytoplasm and the cytoplasmic membrane and on the aggregation current induced by the interaction between membrane-bound proteins. Oscillatory solutions appear via a dynamic instability of the homogenous protein distributions. Moreover, we perform stochastic simulations based on a microscopic description, whereby the probability for each event is calculated according to the corresponding probability in the master equation. Starting from this microscopic description, we derive Langevin equations for the fluctuating protein densities which correspond to the deterministic equations in the limit of vanishing noise. Stochastic simulations justify this deterministic model, showing that oscillations are resistant to the perturbations induced by the stochastic reactions and diffusion. Predictions and assumptions of our theoretical model are compatible with our experimental findings. Altogether, these results enable us to propose further experiments in order to quantitatively compare the different models proposed so far and to test our model with even higher precision. They also point to the necessity of performing such an analysis through single cell measurements.

La problématique de ce doctorat est : comment la structure des réseaux constitue une nouvelle matrice pour la production d’oeuvres artistiques. Pour répondre à cette question, nous commencerons en étudiant l’évolution de la notion de « réseau » de l’antiquité jusqu’à aujourd’hui ; les théories des réseaux concernant leur structure et/ou leur dynamique. Ensuite nous présenterons les applications –de la notion ou des théories de réseaux– tant dans les sciences que dans l’art. Nous listerons et nous analyserons huit types de réseaux et puis nous mentionnerons des oeuvres artistiques qui ont été inspirées par ces types de réseaux et/ou qui utilisent certains (types de) réseaux comme matrice pour leur création. Nous proposerons le projet Lignes Imaginaires, un modèle 3D qui se fonde sur la conception d’un métro de lignes imaginaires, voire de lignes dynamiques et/ou paradoxales qui sont en mouvement, apparaissent/disparaissent, créent de l’infrastructure supplémentaire. L’analyse du métro Lignes Imaginaires dévoile l'importance de la géographie et de la spatialité des réseaux, tandis que leur représentation graphique topologique reste insuffisante pour la représentation précise et pour la compréhension de leur structure (paradoxale). En outre, l’innovation du métro Lignes Imaginaires est que son infrastructure est dynamique et auto-organisée, contrairement aux métros traditionnels où les lignes et leurs itinéraires sont fixes. L’objectif du projet artistique Lignes Imaginaires est de visualiser un concept en créant un métro hors du commun qui pourrait aussi proposer des formes alternatives des réseaux de transports dans le contexte urbain. / This thesis examines and shows ways in which the structure of networks can provide a new matrix for the production of artworks. In order to answer this question we start by studying: the evolution of the term ‘network’ from the ancient times up to nowadays; the theories that refer to network structure or network dynamics. Then we present the applications of these theories into both art and science. We list and analyze eight different types of networks and then we feature artworks which have been inspired by these network types or have used the network structure of a certain type as a matrix for art making. We propose the Imaginary Lines project, a three-dimensional network model which is based on the concept of a metro composed of imaginary lines. More precisely Imaginary Lines metro network encompasses seven paradoxical lines which move, (dis)appear and produce supplementary infrastructure. The Imaginary Lines metro unveils the importance of geography and spatiality, in contrast with topological network graphic representations, which remain insufficient, in terms of utmost accuracy in representation and comprehension of network structure. Additionally, the Imaginary Lines network innovation lays in its infrastructure dynamics as well as in its self-organisation. The objective of the Imaginary Lines artistic project is to visualise a concept by creating an unusual metro, which goes beyond traditional fixed-route transport networks and can support alternative forms of urban transport development.

The subject of this thesis is the generation of spatial temporal structures in living cells. Specifically, we studied the Min-system in the bacterium Escherichia coli. It consists of the MinC, the MinD, and the MinE proteins, which play an important role in the correct selection of the cell division site. The Min-proteins oscillate between the two cell poles and thereby prevent division at these locations. In this way, E. coli divides at the center, producing two daughter cells of equal size, providing them with the complete genetic patrimony. Our goal is to perform a quantitative study, both theoretical and experimental, in order to reveal the mechanism underlying the Min-oscillations. Experimentally, we characterize theMin-system, measuring the temporal period of the oscillations as a function of the cell length, the time-averaged protein distributions, and the in vivo Min-protein mobility by means of different fluorescence microscopy techniques. Theoretically, we discuss a deterministic description based on the exchange of Minproteins between the cytoplasm and the cytoplasmic membrane and on the aggregation current induced by the interaction between membrane-bound proteins. Oscillatory solutions appear via a dynamic instability of the homogenous protein distributions. Moreover, we perform stochastic simulations based on a microscopic description, whereby the probability for each event is calculated according to the corresponding probability in the master equation. Starting from this microscopic description, we derive Langevin equations for the fluctuating protein densities which correspond to the deterministic equations in the limit of vanishing noise. Stochastic simulations justify this deterministic model, showing that oscillations are resistant to the perturbations induced by the stochastic reactions and diffusion. Predictions and assumptions of our theoretical model are compatible with our experimental findings. Altogether, these results enable us to propose further experiments in order to quantitatively compare the different models proposed so far and to test our model with even higher precision. They also point to the necessity of performing such an analysis through single cell measurements.