Academic literature on the topic 'Classification APG'

With the growing size of Application Programming Interfaces (APIs), both API usability and API learning become more challenging. API learning resources are often crucial for helping developers learn an API, but they are distributed across different documents, which makes finding the necessary information more challenging. This work focuses on discovering relevant sections of tutorials for a given API type. We approach this problem by identifying API types in an API tutorial, dividing the tutorial into small fragments and classifying them based on linguistic and structural features. The system we developed can ease information discovery for the developers who need information about a particular API type. Experiments conducted on five tutorials show that our approach is able to discover sections relevant to an API type with 0.79 average precision, 0.73 average recall, and 0.75 average F1 measure when trained and tested on the same tutorial. When trained on four tutorials and tested on a fifth tutorial the average precision is 0.84, average recall is 0.62, and the F1 measure is 0.71.<br>Avec la taille grandissante des interfaces de programmation (API), l'aptitude àl'utilisation ainsi que la facilité d'apprentissage deviennent des préoccupations de premier ordre. La disponibilité de ressources d'apprentissage des API est de grande importance pour parvenir à developer efficacement à partir de différentes sources de documentation. Ce mémoire est consacré au problème de découverte automatique de sections pertinentes contenues dans les tutoriels des API. Nous traitons ce problème en commençant par l'identification du type d'API d'un tutoriel pour ensuite le diviser en fragments qui seront classés d'après leurs propriétés structurelles et linguistiques. Le système que nous avons développé rend le processus de découverte de sections de tutoriel beaucoup plus facile. Une évaluation de notre système a été réalisée avec cinq tutoriels et montre que notre approche peut découvrir des sections pertinentes avec une précision moyenne de 0.79, 0.73 en moyenne de rappel, et 0.75 de mesure moyenne F1 lorsque entraîné ettesté pour le même tutoriel. Lorsqu'entraîné depuis quatre tutoriels et testé dans avec le cinquième, nous obtenons 0.84 de précision moyenne, 0.62 de moyenne de rappel, et finalement 0.71 de mesure F1

Android malware collusion is a new threat model that occurs when multiple Android apps communicate in order to execute an attack. This threat model threatens all Android users' private information and system resource security. Although recent research has made advances in collusion detection and classification, security analysts still do not have robust tools which allow them to definitively identify colluding Android applications. Specifically, in order to determine whether an alert produced by a tool scanning for Android collusion is a true-positive or a false-positive, the analyst must perform manual analysis of the suspected apps, which is both time consuming and prone to human errors. In this thesis, we present a new approach to definitive Android collusion detection and confirmation by rendering inter-component communications between a set of potentially collusive Android applications. Inter-component communications (abbreviated to ICCs), are a feature of the Android framework that allows components from different applications to communicate with one another. Our approach allows Android security analysts to inspect all ICCs within a set of suspicious Android applications and subsequently identify collusive attacks which utilize ICCs. Furthermore, our approach also visualizes all potentially collusive data-flows within each component within a set of apps. This allows analysts to inspect, step-by-step, the the data-flows that are currently used by collusive attacks, or the data-flows that could be used for future collusive attacks. Our tool effectively visualizes the malicious and benign ICCs in sets of proof-of-concept and real-world colluding applications. We conducted a user study which revealed that our approach allows for accurate and efficient identification of true- and false-positive collusive ICCs while still maintaining usability.<br>Master of Science

Master of Science<br>Department of Computing and Information Sciences<br>Doina Caragea<br>Social media affects our daily lives. It is one of the first sources for finding breaking news. In particular, Twitter is one of the popular social media platforms, with around 330 million monthly users. From local events such as Fake Patty's Day to across the world happenings - Twitter gets there first. During a disaster, tweets can be used to post warnings, status of available medical and food supply, emergency personnel, and updates. Users were practically tweeting about the Hurricane Sandy, despite lack of network during the storm. Analysis of these tweets can help monitor the disaster, plan and manage the crisis, and aid in research. In this research, we use the publicly available tweets posted during several disasters and identify the relevant tweets. As the languages in the datasets are different, Bing translation API has been used to detect and translate the tweets. The translations are then, used as training datasets for supervised machine learning algorithms. Supervised learning is the process of learning from a labeled training dataset. This learned classifier can then be used to predict the correct output for any valid input. When trained to more observations, the algorithm improves its predictive performance.

Les signaux d'avertissement que portent les proies toxiques (ou autrement défendues) offrent une opportunité unique de développer une vision intégrative de la diversification biologique. Ces signaux sont soumis à une forte sélection naturelle et sexuelle. D'une part, la stratégie d'échantillonnage utilisée par les prédateurs, caractérisée par l'apprentissage des signaux associés à la toxicité, protège les signaux en forte fréquence dans la communauté de proies. Cette sélection fréquence-dépendante positive favorise l'uniformité des phénotypes et la convergence de signaux entre espèces toxiques (mimétisme mutualiste dit "Müllerien") dans de nombreux taxons. D'autre part, les signaux d'avertissement sont utilisés comme critères de choix entre partenaires sexuels et sont donc soumis à de la sélection sexuelle avec des conséquences importantes pour l'émergence de l'isolement reproducteur et la spéciation. Paradoxalement, malgré une forte sélection naturelle favorisant la convergence, les signaux d'avertissement sont incroyablement diversifiés, à la fois au sein de la même espèce et entre espèces. Cette diversification morphologique est souvent associée à une importante diversification des espèces à l'échelle des clades. Dans cette thèse, j'apporte quelques explications à ce paradoxe et j'affine ainsi notre compréhension de l'effet de la sélection fréquence-dépendante positive et du mutualisme sur la diversification à l'échelle micro- et macro-évolutive. Premièrement, je montre que la stratégie d'échantillonnage des prédateurs peut favoriser la diversification des signaux d'avertissement malgré une sélection fréquence-dépendante positive. Deuxièmement, je dissèque les conditions permettant l'évolution d'un isolement reproducteur stable, nécessaire à la spéciation, dans certaines situations écologiques où les signaux mimétiques sont sous sélection naturelle et sexuelle. Troisièmement, je décris des effets indirect de la sélection fréquence-dépendante sur la diversification à l'échelle macro-évolutive par le biais de contraintes spatiales et de convergences écologiques secondaires<br>The display of warning signals by unpalatable (or otherwise defended) prey provides a wonderful opportunity for establishing an integrative view of biological diversification. Warning signals are known to be under strong natural and sexual selection. On the one hand, the sampling strategy of predators, characterized by a learned avoidance of signals associated with unpalatability, generates natural selection in favour of warning signals in high frequency in the prey community. Such positive frequency-dependent selection favours phenotypic uniformity and causes unpalatable species to converge on common warning signals (mutualistic "Müllerian" mimicry), as seen in a large panel of taxa. On the other hand, warning signals are used as a phenotypic cue for mate choice, generating sexual selection with important consequences for reproductive isolation and speciation. Paradoxically, despite powerful selection favouring phenotypic convergence, warning signals are fantastically diverse, both within and between species, and this morphological diversification is often associated with extensive species diversification at the clade level. In this thesis, I tackle this apparent paradox from the ground up and I thereby refine our understanding of the role of positive frequency-dependent selection and mutualistic interactions for evolutionary diversification at micro- and macroevolutionary scales. First, I show that the predator sampling strategy can favour the emergence of diversity of warning signals despite positive frequency-dependent selection. Second, I dissect the conditions allowing the evolution of strong and stable reproductive isolation, necessary for speciation to occur, in a number of ecological situations where warning signals are under natural and sexual selection. Third, I highlight important indirect effects of frequency-dependent selection on diversification at macro-evolutionary scale via spatial constraints and by-product ecological convergence

Das ehemalige Kornhaus in Zwickau beherbergt eine der modernsten öffentlichen Bibliotheken Sachsens. Vergangenen April war das Gebäude Austragungsort des 22. Länderübergreifenden Seminars zur Kinder- und Jugendbibliotheksarbeit. Vertreter der Bundesländer Hessen, Sachsen, Sachsen-Anhalt und Thüringen haben sich im Rahmen dessen über die Entwicklungen und Potenziale heutiger Medien ausgetauscht.

Exploring upon the phenomena of smart cities, this paper elaborates the potential of crowdsourced data collection in small scale urban quarters. The development of the Data4City (D4C) hyperlocal app – PinCity – is based on the idea of increasing the density of real-time information in urban areas (urban neighborhoods) in order to optimize or create innovative urban services (such as public transportation, garbage collection) or urban planning, thus improving the quality of life of quarter inhabitants as a long-term goal. The main principle of the app is the small-scale implementation, as opposed to top-down smart city approaches worldwide, preferably in a city quarter, or a community, which can be subsequently scaled and interlaced to other parts of the city.

This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.<br>Thesis: M. Eng., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2019<br>Cataloged from student-submitted PDF version of thesis.<br>Includes bibliographical references (pages 129-131).<br>As machine learning permeates our society and manifests itself through commonplace technologies such as autonomous vehicles, facial recognition, and online store recommendations, it is necessary that the increasing number of people who rely on these tools understand how they work. As such, we need to develop effective tools and curricula for introducing machine learning to novices. My work focuses on teaching core machine learning concepts with image classification, one of the most basic and widespread examples of machine learning. I built a web interface that allows users to train and test personalized image classification models on pictures taken with their computers--webcams. Furthermore, I built an extension for MIT App Inventor, a platform for building mobile applications using a blocks-based programming language, that allows users to use the models they built in the web interface to classify objects in their mobile applications. Finally, I created high school level curricula for workshops based on using the aforementioned interface and App Inventor extension, and ran the workshops with two classes of high school students from Boston Latin Academy. My findings indicate that high school students with no machine learning background are able to learn and understand general concepts and applications of machine learning through hands-on, non-technical activities, as well as successfully utilize models they built for personal use.<br>by Danny Tang.<br>M. Eng.<br>M.Eng. Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science

Gerade in Großveranstaltungen geht die Interaktion zwischen Lehrenden und Lernenden häufig verloren. So genannte Classroom Response Systems (CRS) knüpfen an dieser fehlenden Aktivierung und Interaktion an, indem sie als technisches Hilfsmittel die klassischen Interaktionsformen wie Fragen oder Abstimmungen auch in Großveranstaltungen ermöglichen. Ein solches Feedbacksystem wurde mittels der myTU-App ohne kostenintensive Anschaffungen realisiert. Die Entwickler der myTU-App arbeiten stetig an der Weiterentwicklung und Verbesserung der App. Dafür ist es wichtig, sowohl die Bedürfnisse und Wünsche der Nutzer/innen, als auch die Ablehnungsgründe derer, die die App nicht nutzen, zu kennen. Aus diesem Grund wurde eine Evaluation zur myTU-App durchgeführt.