<p> Determining the optimal size of a neural network is complicated. Neural networks, with many free parameters, can be used to solve very complex problems. However, these neural networks are susceptible to overfitting. BCAP (Brantley-Clark Artificial Neural Network Pruning Technique) addresses overfitting by combining duplicate neurons in a neural network hidden layer, thereby forcing the network to learn more distinct features. We compare hidden units using the cosine similarity, and combine those that are similar with each other within a threshold &epsiv;. By doing so the co-adaption of the neurons in the network is reduced because hidden units that are highly correlated (i.e. similar) are combined. In this paper we show evidence that BCAP is successful in reducing network size while maintaining accuracy, or improving accuracy of neural networks during and after training.</p>

<p> Human disease is complex. However, the explosion of biomedical data is providing new opportunities to improve our understanding. My dissertation focused on how to harness the biodata revolution. Broadly, I addressed three questions: how to integrate data, how to extract insights from data, and how to make science more open. </p><p> To integrate data, we pioneered the hetnet&mdash;a network with multiple node and relationship types. After several preludes, we released Hetionet v1.0, which contains 2,250,197 relationships of 24 types. Hetionet encodes the collective knowledge produced by millions of studies over the last half century. </p><p> To extract insights from data, we developed a machine learning approach for hetnets. In order to predict the probability that an unknown relationship exists, our algorithm identifies influential network patterns. We used the approach to prioritize disease&mdash;gene associations and drug repurposing opportunities. By evaluating our predictions on withheld knowledge, we demonstrated the systematic success of our method. </p><p> After encountering friction that interfered with data integration and rapid communication, I began looking at how to make science more open. The quest led me to explore realtime open notebook science and expose publishing delays at journals as well as the problematic licensing of publicly-funded research data.</p>

<p> Scientific datasets associated with a research project can proliferate over time as a result of activities such as sharing datasets among collaborators, extending existing datasets with new measurements, and extracting subsets of data for analysis. As such datasets begin to accumulate, it becomes increasingly difficult for a scientist to keep track of their derivation history, which complicates data sharing, provenance tracking, and scientific reproducibility. Understanding what relationships exist between datasets can help scientists recall their original derivation history. For instance, if dataset <i>A</i> is contained in dataset <i>B,</i> then the connection between <i>A</i> and <i>B</i> could be that <i>A</i> was extended to create <i>B.</i> </p><p> We present a relationship-identification methodology as a solution to this problem. To examine the feasibility of our approach, we articulated a set of relevant relationships, developed algorithms for efficient discovery of these relationships, and organized these algorithms into a new system called ReConnect to assist scientists in relationship discovery. We also evaluated existing alternative approaches that rely on flagging differences between two spreadsheets and found that they were impractical for many relationship-discovery tasks. Additionally, we conducted a user study, which showed that relationships do occur in real-world spreadsheets, and that ReConnect can improve scientists' ability to detect such relationships between datasets. </p><p> The promising results of ReConnect's evaluation encouraged us to explore a more automated approach for relationship discovery. In this dissertation, we introduce an automated end-to-end prototype system, ReDiscover, that identifies, from a collection of datasets, the pairs that are most likely related, and the relationship between them. Our experimental results demonstrate the overall effectiveness of ReDiscover in predicting relationships in a scientist's or a small group of researchers' collections of datasets, and the sensitivity of the overall system to the performance of its various components.</p>

No description available.

published_or_final_version / Computer Science / Doctoral / Doctor of Philosophy

<p> The purpose of this study is to explore the potential threats that are posed uniquely by cyber terrorism. While traditional terrorism has affected governmental policy, and inflicted physical damage to people and infrastructure across the world, computers and the Internet can allow for attacks as well. As terrorism groups begin to adapt to, and take advantage of - cyber tools and capabilities, the threat they pose will grow accordingly. While a terrorist is not able to directly kill people with cyber tools, the mayhem or social disruption that such attacks can cause, fit well with these organization's objectives. The anonymity of cyber space permits terrorist groups to plan and execute attacks without being identified immediately, if ever. In addition, the targets of cyber terrorists are often under prepared, and fairly vulnerable to various forms of cyber attacks. While these organizations may be aware of the risk posed by failing to adequately address cyber security deficiencies, their solutions are likely not sufficient to truly prevent cyber terrorism attacks. In order to discover technological advancements, efficient cyber security, and generally spread awareness on the subject, this study attempts to highlights existing threats, as well as an overview of what can be done to mitigate them. </p><p> Keywords: Intelligence, Cyber security, Professor Albert Orbanati</p>

<p> Decision making under demand uncertainty, a top priority task, has remained as the most challenging problem to many manufacturing leaders due to lack of sufficient information to establish supply chain management (SCM) standard policies. The problem was that business performance could be impeded because optimization models of existing SCM systems lacked appropriate control mechanisms to optimize inventory levels and reduce the bullwhip effect. The purpose of this quantitative correlational study was to investigate the extent to which SCM control mechanisms predict optimized inventory levels (OPT) and reduced bullwhip effect (BWE) based on the perceptions of supply chain (SC) senior-level managers of medium-size and large manufacturing firms in the United States. Model predictive control-based inventory optimization (MPC), internal model control-based inventory optimization (IMC), postponement (POS), and collaboration (COL) were used as predictor variables, and SCM performance was the criterion variables as measured by OPT and BWE. A survey was used to collect data from SC senior-level managers. Regression analysis resulted in two significant regression models for OPT and BWE that explained 61% and 49.7 % of the variance respectively for OPT (<i>p</i> &lt; .05) and BWE (<i>p</i> &lt; .05). As a result, both null hypotheses 1 and 2 were rejected, and support existed for the alternative hypotheses 1 and 2. Practical recommendations included use of MPC to optimize inventory levels, use of POS and COL strategies to reduce the bullwhip effect and optimize inventory levels, and to combine IMC, MPC, POS, and COL to synergistically reduce the bullwhip effect and optimize inventory levels. Recommendations for future research included a replicate quantitative correlation study with expansion to international manufacturing firms, a quantitative structural equation modeling study to examine relative strength and causal relationships among variables, a quantitative meta-analysis study to critically examine the findings of the study across other studies, a quantitative experimental study to further scrutinize the significant relationships between OPT and BWE, and a quantitative experimental study of archival data to reduce self-selection and self-reporting sampling biases.</p>

No description available.

Ce mémoire présente un bilan des travaux de recherches que j'ai dirigé au CRAI (Centre<br />de Recherche en Architecture et Ingénierie) depuis environ dix ans. Avant de développer<br />les deux grandes problématiques qui ont structuré ces années, je voudrais remettre<br />rapidement en perspective mon travail global de recherche dans le vaste mais encore très<br />récent domaine de la recherche en architecture.<br />1.1. LA RECHERCHE EN ARCHITECTURE, QUELS<br />SYSTÈMES DE CONNAISSANCES ?<br />Jusqu'au milieu des années 1960, l'architecture est avant tout un art, c'est-à-dire une<br />pratique reposant largement sur l'expérience, fondée sur des grands traités et infléchie par<br />quelques maîtres.<br />Au niveau le plus empirique, le terme de recherche ne désigne rien d'autre que la pratique<br />quotidienne du concepteur en situation de création pour inventer un dispositif spatial en<br />réponse à des exigences explicites ou implicites.<br />Comme le musicien, le peintre ou le sculpteur, l'architecte est une sorte de « chercheur de<br />formes », un inventeur de langages, un constructeur de sens.<br />Dans cet univers de la recherche-action, le concepteur utilise et produit des connaissances<br />qui relèvent plus d'un savoir-faire que d'un véritable savoir. À cette recherche, tout ancrée<br />dans la projetation, nous donnons le nom de recherche architectonique.<br />Dans ce dispositif, le terme d'architecture désigne donc à la fois un domaine de production<br />(essentiellement les édifices et les espaces urbains) et un savoir-faire largement issu et<br />opératoire dans la pratique du projet. Mais pour être efficace, cette économie de la<br />production architecturale a besoin d'un discours régulateur et légitimateur destiné à<br />permettre des choix tout au long du processus de conception, évaluer les solutions<br />partielles ou globales proposées et juger la qualité des œuvres produites. C'est le rôle des<br />traités, parfois nommés théories.<br />Les traités d'architecture de Vitruve à Le Corbusier bien que théoriques par le recul qu'ils<br />prenaient par rapport aux pratiques de leurs époques respectives ne peuvent pas<br />cependant être considérés comme des théories au sens scientifique du terme. Ce<br />deuxième type de connaissances, auxquelles nous donnons le nom de doctrine est un<br />système de préceptes et d'axiomes transmissibles, orientés par l'action vers l'action. Par<br />différence avec les connaissances souvent singulières et peu externalisables des savoir-

<p>Privat annonsering har flyttats från dagspress till nätet vilket har bidragit till bättre sökfunktioner och rikare beskrivning. Då tiden vi spenderar på nätet ökar så öppnar det för att kunna kommunicera i realtid med användare som är online. Tyvärr finns det ingen bra tjänst idag för att kunna kommunicera på något annats sätt än telefon och e-mail.</p><p>För att skapa ett rikare sätt att kommunicera så skapade jag denna tjänst som gör det möjligt för användare att föra olika konversationer. För att öka säkerhet mellan både säljare och köpare har det skapats betygsättning och att användare registrerar personuppgifter för kontroll av den andra parten. Detta arbete har varit av stort eget intresse för att öka kommunikationsmöjligheterna mellan människor.</p><p>De tekniker som jag använt har bestått av PHP ramverket Codeigniter som bygger på MVC. För presentation har jag jobbat mycket med jQuery och CSS för att skapa de effekter som jag velat åstadkomma. För lagning har jag använt MySQL. Resultatet blev en sida som är en annonssida med den funktionalitet som en Community har.</p>