Submitted by Fabricia Fialho Reginato (fabriciar) on 2015-08-26T23:54:22Z No. of bitstreams: 1 TiagoWiedemann.pdf: 5635533 bytes, checksum: c0d3805abbcaf56aa36da4d8422457b6 (MD5) / Made available in DSpace on 2015-08-26T23:54:22Z (GMT). No. of bitstreams: 1 TiagoWiedemann.pdf: 5635533 bytes, checksum: c0d3805abbcaf56aa36da4d8422457b6 (MD5) Previous issue date: 2014 / CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / CNPQ – Conselho Nacional de Desenvolvimento Científico e Tecnológico / A Computação Ubíqua, que estuda formas de integrar a tecnologia ao cotidiano das pessoas, é uma área que vem crescendo nos últimos anos, especialmente devido ao desenvolvimento de tecnologias como a computação móvel. Um dos aspectos fundamentais para o desenvolvimento deste tipo de aplicação é a questão da Sensibilidade ao Contexto, que permite a uma aplicação adaptar o seu funcionamento conforme a situação na qual o usuário se encontra no momento. Com esta finalidade, diversos autores apresentaram definições formais sobre o que é um contexto e como representá-lo. A partir desta formalização começaram a ser desenvolvidas técnicas para análise de dados contextuais que propunham a realização de predições e inferências, entre outras análises. Esta dissertação especifica um framework denominado SIMCOP (SIMilar Context Path) para a realização da análise de similaridade entre sequências de contextos visitados por uma entidade. Este tipo de análise permite a identificação de contextos semelhantes com a intenção de prover funcionalidades como a recomendação de entidades e/ou contextos, a classificação de entidades e a predição de contextos. Um protótipo do framework foi implementado, e a partir dele foram desenvolvidas duas aplicações de recomendação, uma delas por um desenvolvedor independente, através do qual foi possível avaliar a eficácia do framework. Com o desenvolvimento desta pesquisa comprovou-se, conforme demonstrado nas avaliações realizadas, que a análise de similaridade de contextos pode ser útil em outras áreas além da computação ubíqua, como a mineração de dados e os sistemas de filtragem colaborativa, entre outras áreas, onde qualquer conjunto de dados que puder ser descrito na forma de um contexto, poderá ser analisado através das técnicas de análise de similaridade implementadas pelo framework. / The Ubiquitous Computing, that studies the ways to integrate technology into the people’s everyday life, is an area that has been growing in recent years, especially due to the development of technologies such as mobile computing. A key for the development of this type of application is the issue of context awareness, which enables an application to self adapt to the situation in which the user is currently on. To make this possible, it was necessary to formally define what is a context and how to represent it . From this formalization, techniques for analyzing contextual data have been proposed for development of functions as predictions or inferences. This paper specifies a framework called SIMCOP (SIMilar Context Path ) for performing the analysis of similarity between sequences of contexts visited by an entity. This type of analysis enables the identification of similar contexts with the intention to provide features such as the recommendation of entities and contexts, the entities classification and the prediction of contexts. The development of this research shows that the contexts similarity analysis can be useful in other areas further the ubiquitous computing, such as data mining and collaborative filtering systems. Any data type that can be described as a context, can be analyzed through the techniques of similarity analysis implemented by the framework, as demonstrated in the assessments.

Les cadriciels et les bibliothèques sont indispensables aux systèmes logiciels d'aujourd'hui. Quand ils évoluent, il est souvent fastidieux et coûteux pour les développeurs de faire la mise à jour de leur code. Par conséquent, des approches ont été proposées pour aider les développeurs à migrer leur code. Généralement, ces approches ne peuvent identifier automatiquement les règles de modification une-remplacée-par-plusieurs méthodes et plusieurs-remplacées-par-une méthode. De plus, elles font souvent un compromis entre rappel et précision dans leur résultats en utilisant un ou plusieurs seuils expérimentaux. Nous présentons AURA (AUtomatic change Rule Assistant), une nouvelle approche hybride qui combine call dependency analysis et text similarity analysis pour surmonter ces limitations. Nous avons implanté AURA en Java et comparé ses résultats sur cinq cadriciels avec trois approches précédentes par Dagenais et Robillard, M. Kim et al., et Schäfer et al. Les résultats de cette comparaison montrent que, en moyenne, le rappel de AURA est 53,07% plus que celui des autre approches avec une précision similaire (0,10% en moins). / Software frameworks and libraries are indispensable to today's software systems. As they evolve, it is often time-consuming for developers to keep their code up-to-date. Approaches have been proposed to facilitate this. Usually, these approaches cannot automatically identify change rules for one-replaced-by-many and many-replaced-by-one methods, and they trade off recall for higher precision using one or more experimentally-evaluated thresholds. We introduce AURA (AUtomatic change Rule Assistant), a novel hybrid approach that combines call dependency and text similarity analyses to overcome these limitations. We implement it in a Java system and compare it on five frameworks with three previous approaches by Dagenais and Robillard, M. Kim et al., and Schäfer et al. The comparison shows that, on average, the recall of AURA is 53.07% higher while its precision is similar (0.10% lower).

Made available in DSpace on 2014-06-11T19:35:43Z (GMT). No. of bitstreams: 0 Previous issue date: 2009-04-06Bitstream added on 2014-06-13T19:05:56Z : No. of bitstreams: 1 santos_tg_dr_rcla.pdf: 2207082 bytes, checksum: cd6f4211c558cc08a16cda8ec6665379 (MD5) / Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / Estudamos a riqueza, composição taxonômica e os padrões de distribuição espaciais e temporais de taxocenoses de anuros do Parque Estadual Morro do Diabo (PEMD), o maior remanescente de Floresta Estacional Semidecídua no estado de São Paulo, Brasil. Registramos 28 espécies de anuros (Apêndice I) de setembro de 2005 a março de 2007, que compreenderam um misto de espécies de Mata Atlântica, do Cerrado e de formas amplamente distribuídas na América do Sul, geralmente consideradas tolerantes a modificações antropogênicas. A baixa riqueza de espécies e de modos reprodutivos, a predominância de espécies habitatgeneralistas e a alta similaridade de espécies de anuros com áreas de Cerrado podem ser explicadas pela sazonalidade climática da área estudada (estação seca pronunciada), além da grande distância em relação a centros de diversificação de anuros, como as montanhas costeiras da Floresta Atlântica úmida. Chuva e fotoperíodo explicaram aproximadamente 77% da atividade de vocalização de toda a taxocenose, enquanto somente a chuva e o fotoperíodo explicaram a temporada de vocalização em ambientes temporários e permanentes, respectivamente. Registramos alta sobreposição na temporada de vocalização dos machos, mas segregação na fase larval. A distribuição das espécies de anuros entre sítios de reprodução (Apêndice II) diferiu da esperada pelo acaso e compreendeu três taxocenoses distintas de anuros que foram explicadas pelo conjunto de variáveis ambientais de riachos permanentes, represas permanentes e poças temporárias. Registramos que 19 espécies de anuros (aproximadamente 83% da riqueza total de espécies registradas nos corpos d’água monitorados) foram indicadoras da heterogeneidade ambiental: três espécies indicaram riachos permanentes, quatro indicaram represas permanentes e 12 espécies indicaram poças temporárias... / We studied richness, composition, and patterns of temporal and spatial distributions of anuran assemblages of Morro do Diabo State Park (MDSP), the major remnant of Mesophytic Semideciduous Forest (MSF) in the state of São Paulo, Brazil. From September 2005 to March 2007 we recorded 28 anuran species (Appendix I), comprising a mix of Atlantic, Cerrado, and South American widespread species, usually considered tolerant to anthropic modifications. The low richness of species and reproductive modes, the predominance of habitat generalist species, and the high similarity with Cerrado areas can be explained by climatic seasonality of the studied area (pronounced dry season), besides its large distance in relation to centers of anuran diversification, such as coastal mountains of the wet Atlantic Forest. Rainfall and photoperiod explained about 77% of calling activity of the whole assemblage, while rainfall alone in temporary habitats and photoperiod in permanent ones explained the calling season. We recorded high temporal overlap for calling males, but segregation for tadpoles. Spatial distribution of anuran species among breeding sites of the MDSP (Appendix II) differed of expected by chance and comprised three distinct anuran assemblages that were explained by the suite environmental variables of permanent streams, permanent dams, and temporary ponds. We recorded that 19 species (about 83% of total anuran species recorded in monitored sites) were indicators of environmental heterogeneity: three anuran species indicated permanent streams, four indicated permanents dams, and 12 anuran species indicated temporary ponds. Regarding to micro-spatial distribution of anuran species at two temporary ponds of MDSP, we recorded that males of most pairs of species (96%) used distinct sites for calling activities. The best combination of variables discriminating anuran species regarding male... (Complete abstract click electronic access below)

Submitted by Johnny Rodrigues (johnnyrodrigues@ufcg.edu.br) on 2017-12-11T13:08:51Z No. of bitstreams: 1 DANIELLE CHAVES DE MEDEIROS - DISSERTAÇÃO PPGCC 2014.pdf: 1794393 bytes, checksum: 843f24f727881568e52ff59c12122403 (MD5) / Made available in DSpace on 2017-12-11T13:08:51Z (GMT). No. of bitstreams: 1 DANIELLE CHAVES DE MEDEIROS - DISSERTAÇÃO PPGCC 2014.pdf: 1794393 bytes, checksum: 843f24f727881568e52ff59c12122403 (MD5) Previous issue date: 2014-12 / A seleção de critérios para a análise de informações durante o processo de avaliação da participação em fóruns de discussão de cursos de Educação a Distância (EaD) é um grande desafio. São muitas as variáveis que devem ser consideradas neste processo, além da subjetividade inerente à análise realizada pelo docente, passível de erro humano. Os docentes geralmente não possuem a seu dispor todos os recursos necessários, se tornando necessário o uso de uma metodologia ou ferramenta que o auxilie no processo de avaliação. Diante desta demanda e, a partir de um estudo dos principais indicadores qualitativos/quantitativos utilizados pelos professores de EaD, foi desenvolvido um arcabouço para a análise da participação dos alunos em fóruns. O objetivo deste arcabouço é servir de apoio à tomada de decisão do professor, fornecendo um mecanismo mais efetivo para a mensuração da quantidade e da qualidade das interações, passível de adaptação à metodologia tradicional adotada por cada docente. A validação deste arcabouço deu-se a partir da administração de questionários para a sondagem da opinião de docentes atuantes na área de ensino a distância, assim como pela realização de estudos de caso envolvendo a avaliação da acurácia de instâncias do arcabouço para o cálculo da nota de participação de alunos. Foi desenvolvido um Sistema Especialista (SE) para o processamento dos dados, com o uso de funções de similaridade para realizar, de forma semiautomática, a avaliação do conteúdo das mensagens dos alunos. Assim, as notas de participação calculadas foram confrontadas com as notas atribuídas pelo docente utilizando a abordagem tradicional. Os resultados obtidos demonstraram que, em três das cinco turmas observadas, não foi possível verificar a existência de diferenças estatísticas significativas entre o desempenho das abordagens estudadas. Um estudo da acurácia e correlação revela que, em todos os casos analisados, há uma forte relação entre os dados e o erro médio encontrado foi inferior a 3%, demonstrando a aplicabilidade do arcabouço ao contexto da avaliação da participação em fóruns. / The selection of criteria for the information analysis during the participation evaluation process, in discussion forums of distance courses, is a major challenge. There are many variables to consider in this process, in addition to the subjectivity inherent in the analysis carried out by the instructor, which is subject to human error. Instructors, generally, do not have at their disposal all the resources necessary, thus, the use of methodologies or tools that can help them with this process are necessary. Facing this demand, and after performing a study of the major qualitative/quantitative indicators used by distance education teachers, we developed a framework for the analysis of the student participation in the forums. The aim of this framework is to support the decision-making process, by providing a more effective mechanism to measure the quantity and quality of interactions, capable of adjusting itself to the traditional methodology adopted by each teacher. The validation of this framework was performed by the administration of questionnaires that surveyed the opinion of active teachers in the distance learning area, and by the execution of case studies involving the assessment of the accuracy of instances of this framework for calculating the participation grade of students. The study involved the development of an Expert System, for the treatment and processing of the data, using similarity functions to perform, semi-automatically, the assessment of the content of the students' messages. Thus, it was possible to confront the calculated participation grades with the grades assigned by the teacher using the traditional approach. The results showed that, in three out of the five classes observed, it was not possible to verify the existence of statistically significant differences between the performance of both the approaches studied. A study of the accuracy and correlation shows that, in all the cases analyzed, there is a strong relationship between the data, and the average error was less than 3%, demonstrating the applicability of the proposed framework to the assessment of student participation in forums.

CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Melastomataceae Juss, a maior família da ordem Myrtales, apresenta ampla distribuição pantropical. A família é uma das maiores da flora brasileira, somando cerca de 150 gêneros e 4.500 espécies. Melastomataceae é dividida em 11 tribos, sendo Melastomeae Bartl., Miconieae DC. e Microlicieae Naudin as tribos mais importantes para o Brasil, onde podem ser encontradas cerca de 1/3 de suas espécies. Microlicia D. Don, o maior gênero da tribo Microliciae, é um dos gêneros mais representativos, com cerca de 120 espécies, concentradas principalmente nos campos rupestres do Brasil Central. Embora atinjam sua maior diversidade nos campos rupestres, algumas espécies de Microlicia têm ampla distribuição, podendo ser encontradas em outras fitofisionomias, como cerrado, veredas, campos úmidos e campos hidromórficos. Microlicia é considerado um gênero problemático entre as Melastomataceae, sendo suas espécies reconhecidas apenas pela combinação de diferentes características, o que torna difícil a identificação precisa de várias delas. Numa tentativa de elucidar a delimitação dos taxa mais problemáticos, um grande esforço tem sido dedicado ao estudo da anatomia foliar de espécies do gênero na última década, e hoje cerca de 1/3 das espécies já tiveram a anatomia foliar descrita. As características mais comuns para o gênero são: epiderme uniestratificada, folhas anfiestomáticas, presença de apêndices epidérmicos variados, mesofilo isobilateral ou homogêneo, e presença de compostos fenólicos nos tecidos da folha, principalmente nas células do parênquima paliçádico. No entanto, apesar das várias características em comum, as folhas de Microlicia exibem também características anatômicas diferentes, que poderiam auxiliar na identificação dos taxa de reconhecimento mais impreciso. Considerando o grande número de espécies de Microlicia com a anatomia foliar descrita, o polimorfismo relatado para várias espécies do gênero e o possível potencial plástico da folha, que as espécies de ampla ocorrência poderiam apresentar, o presente estudo teve como objetivos gerais: a) descrever a anatomia foliar de mais nove espécies do gênero, ocorrentes nos campos rupestres; b) descrever a anatomia foliar de quatro espécies ocorrentes em diferentes ambientes, além dos campos rupestres e c) através da análise de similaridade, realizada com base nas características anatômicas da folha de todas as Microlicia estudadas até o momento, analisar os grupos de espécies semelhantes formados e identificar caracteres que propiciem a formação dos mesmos. / Melastomataceae Juss, the largest family of the order Myrtales, is widely distributed pantropical. The family is one of the largest of the Brazilian flora, with about 150 genera and 4,500 species. Melastomataceae is divided into 11 tribes, with Melastomeae Bartl., Miconieae DC. and Microlicieae Naudin the most important tribes in Brazil, and in Brazil can be found about 1/3 of its species. Microlicia D. Don the largest genus of Microliciae tribe and one of the most representative genera, with about 120 species, mainly concentrated to the “campos rupestres” in the Central Brazil. Although Microlicia reaching its greatest diversity in the Brazilian “campos rupestres”, some species are widely distributed and can be found in other vegetation types such as cerrado, paths, swamps and hydromorphic fields. Microlicia is considered a problematic genus among the Melastomataceae, and their species are recognized only by combining different traits, making it difficult to accurately identify several of them. In an attempt to help the delimitation of the most problematic species, in the last decade a great effort has been devoted to the study of Microlicia leaf anatomy and today around 1/3 of these species have had the leaf anatomy described. The most common features are: unisseriate epidermis, amphistomatic leaves, presence of trichomes and emergences, isobilateral or homogeneous mesophyll and presence of phenolic compounds in leaf tissues, particularly in the palisade parenchyma cells. However, despite the many similar features, Microlicia leaves also exhibit different anatomical characteristics that could help identify the most problematic species. Considering the large number of Microlicia species with the leaf anatomy known, the polymorphism already reported for several species of the genus, and the possible plastic potential of the leaf anatomy that the species of widely spread could present, the goals of this work were: a) analyze the leaf anatomy of nine Microlicia from “campos rupestres”; b) analyze and interpret the leaf anatomy of four species occurring in others environments, beyond the “campos rupestres” and c) through similarity analysis performed using the anatomical leaf traits of the all Microlia studied to date, analyze the groups of similar species formed and identify characters that favor the formation of the same. / Dissertação (Mestrado)

Le traitement auditif dans le cerveau humain et dans les systèmes informatiques consiste en une cascade de transformations représentationnelles qui extraient et réorganisent les informations pertinentes pour permettre l'exécution des tâches. Cette thèse s'intéresse à la nature des représentations acoustiques et aux principes de conception et d'apprentissage qui soutiennent leur développement. Les objectifs scientifiques sont de caractériser et de comparer les représentations auditives dans les réseaux de neurones convolutionnels profonds (CNN) et la voie auditive humaine. Ce travail soulève plusieurs questions méta-scientifiques sur la nature du progrès scientifique, qui sont également considérées. L'introduction passe en revue les connaissances actuelles sur la voie auditive des mammifères et présente les concepts pertinents de l'apprentissage profond. Le premier article soutient que les questions philosophiques les plus pressantes à l'intersection de l'intelligence artificielle et biologique concernent finalement la définition des phénomènes à expliquer et ce qui constitue des explications valables de tels phénomènes. Je surligne les théories pertinentes de l'explication scientifique que j’espére fourniront un échafaudage pour de futures discussions. L'article 2 teste un modèle populaire de cortex auditif basé sur des modulations spectro-temporelles. Nous constatons qu'un modèle linéaire entraîné uniquement sur les réponses BOLD aux ondulations dynamiques simples (contenant seulement une fréquence fondamentale, un taux de modulation temporelle et une échelle spectrale) peut se généraliser pour prédire les réponses aux mélanges de deux ondulations dynamiques. Le troisième article caractérise la spécificité linguistique des couches CNN et explore l'effet de l'entraînement figé et des poids aléatoires. Nous avons observé trois régions distinctes de transférabilité: (1) les deux premières couches étaient entièrement transférables, (2) les couches 2 à 8 étaient également hautement transférables, mais nous avons trouvé évidence de spécificité de la langue, (3) les couches suivantes entièrement connectées étaient plus spécifiques à la langue mais pouvaient être adaptées sur la langue cible. Dans l'article 4, nous utilisons l'analyse de similarité pour constater que la performance supérieure de l'entraînement figé obtenues à l'article 3 peuvent être attribuées aux différences de représentation dans l'avant-dernière couche: la deuxième couche entièrement connectée. Nous analysons également les réseaux aléatoires de l'article 3, dont nous concluons que la forme représentationnelle est doublement contrainte par l'architecture et la forme de l'entrée et de la cible. Pour tester si les CNN acoustiques apprennent une hiérarchie de représentation similaire à celle du système auditif humain, le cinquième article compare l'activité des réseaux «freeze trained» de l'article 3 à l'activité IRMf 7T dans l'ensemble du système auditif humain. Nous ne trouvons aucune évidence d'une hiérarchie de représentation partagée et constatons plutôt que tous nos régions auditifs étaient les plus similaires à la première couche entièrement connectée. Enfin, le chapitre de discussion passe en revue les mérites et les limites d'une approche d'apprentissage profond aux neurosciences dans un cadre de comparaison de modèles. Ensemble, ces travaux contribuent à l'entreprise naissante de modélisation du système auditif avec des réseaux de neurones et constituent un petit pas vers une science unifiée de l'intelligence qui étudie les phénomènes qui se manifestent dans l'intelligence biologique et artificielle. / Auditory processing in the human brain and in contemporary machine hearing systems consists of a cascade of representational transformations that extract and reorganize relevant information to enable task performance. This thesis is concerned with the nature of acoustic representations and the network design and learning principles that support their development. The primary scientific goals are to characterize and compare auditory representations in deep convolutional neural networks (CNNs) and the human auditory pathway. This work prompts several meta-scientific questions about the nature of scientific progress, which are also considered. The introduction reviews what is currently known about the mammalian auditory pathway and introduces the relevant concepts in deep learning.The first article argues that the most pressing philosophical questions at the intersection of artificial and biological intelligence are ultimately concerned with defining the phenomena to be explained and with what constitute valid explanations of such phenomena. I highlight relevant theories of scientific explanation which we hope will provide scaffolding for future discussion. Article 2 tests a popular model of auditory cortex based on frequency-specific spectrotemporal modulations. We find that a linear model trained only on BOLD responses to simple dynamic ripples (containing only one fundamental frequency, temporal modulation rate, and spectral scale) can generalize to predict responses to mixtures of two dynamic ripples. Both the third and fourth article investigate how CNN representations are affected by various aspects of training. The third article characterizes the language specificity of CNN layers and explores the effect of freeze training and random weights. We observed three distinct regions of transferability: (1) the first two layers were entirely transferable between languages, (2) layers 2--8 were also highly transferable but we found some evidence of language specificity, (3) the subsequent fully connected layers were more language specific but could be successfully finetuned to the target language. In Article 4, we use similarity analysis to find that the superior performance of freeze training achieved in Article 3 can be largely attributed to representational differences in the penultimate layer: the second fully connected layer. We also analyze the random networks from Article 3, from which we conclude that representational form is doubly constrained by architecture and the form of the input and target. To test whether acoustic CNNs learn a similar representational hierarchy as that of the human auditory system, the fifth article presents a similarity analysis to compare the activity of the freeze trained networks from Article 3 to 7T fMRI activity throughout the human auditory system. We find no evidence of a shared representational hierarchy and instead find that all of our auditory regions were most similar to the first fully connected layer. Finally, the discussion chapter reviews the merits and limitations of a deep learning approach to neuroscience in a model comparison framework. Together, these works contribute to the nascent enterprise of modeling the auditory system with neural networks and constitute a small step towards a unified science of intelligence that studies the phenomena that are exhibited in both biological and artificial intelligence.

Making complex medical decisions is becoming an increasingly challenging task due to the growing amount of available evidence to consider and the higher demand for personalized treatment and patient care. IT systems for the provision of clinical decision support (CDS) can provide sustainable relief if decisions are automatically evaluated and processed. In this paper, we propose an approach for quantifying similarity between new and previously recorded medical cases to enable significant knowledge transfer for reasoning tasks on a patient-level. Methodologically, 102 medical cases with oropharyngeal carcinoma were analyzed retrospectively. Based on independent disease characteristics, patient-specific data vectors including relevant information entities for primary and adjuvant treatment decisions were created. Utilizing the ϕK correlation coefficient as the methodological foundation of our approach, we were able to determine the predictive impact of each characteristic, thus enabling significant reduction of the feature space to allow for further analysis of the intra-variable distances between the respective feature states. The results revealed a significant feature-space reduction from initially 19 down to only 6 diagnostic variables (ϕK correlation coefficient ≥ 0.3, ϕK significance test ≥ 2.5) for the primary and 7 variables (from initially 14) for the adjuvant treatment setting. Further investigation on the resulting characteristics showed a non-linear behavior in relation to the corresponding distances on intra-variable level. Through the implementation of a 10-fold cross-validation procedure, we were further able to identify 8 (primary treatment) matching cases with an evaluation score of 1.0 and 9 (adjuvant treatment) matching cases with an evaluation score of 0.957 based on their shared treatment procedure as the endpoint for similarity definition. Based on those promising results, we conclude that our proposed method for using data-driven similarity measures for application in medical decision-making is able to offer valuable assistance for physicians. Furthermore, we consider our approach as universal in regard to other clinical use-cases, which would allow for an easy-to-implement adaptation for a range of further medical decision-making scenarios.

No description available.

Supervised learning requires labeled data which is cumbersome to produce, making it costly and time-consuming. SimCLR is a self-supervising framework that uses data augmentations to learn without labels. This thesis investigates how well cropping and color distorting augmentations work for two datasets, MPI3D and Causal3DIdent. The representations learned are evaluated using representation similarity analysis. The data augmentations were meant to make the model learn invariant representations of the object shape in the images regarding it as content while ignoring unnecessary features and regarding them as style. As a result, 8 models were created, models A-H. A and E were trained using supervised learning as a benchmark for the remaining self-supervised models. B and C learned invariant features of style instead of learning invariant representations of shape. Model D learned invariant representations of shape. Although, it also regarded style-related factors as content. Model F, G, and H managed to learn invariant representations of shape with varying intensities while regarding the rest of the features as style. The conclusion was that models can learn invariant representations of features related to content using self-supervised learning with the chosen augmentations. However, the augmentation settings must be suitable for the dataset. / Övervakad maskininlärning kräver annoterad data, vilket är dyrt och tidskrävande att producera. SimCLR är ett självövervakande maskininlärningsramverk som använder datamodifieringar för att lära sig utan annoteringar. Detta examensarbete utvärderar hur väl beskärning och färgförvrängande datamodifieringar fungerar för två dataset, MPI3D och Causal3DIdent. De inlärda representationerna utvärderas med hjälp av representativ likhetsanalys. Syftet med examensarbetet var att få de självövervakande maskininlärningsmodellerna att lära sig oföränderliga representationer av objektet i bilderna. Meningen med datamodifieringarna var att påverka modellens lärande så att modellen tolkar objektets form som relevant innehåll, men resterande egenskaper som icke-relevant innehåll. Åtta modeller skapades (A-H). A och E tränades med övervakad inlärning och användes som riktmärke för de självövervakade modellerna. B och C lärde sig oföränderliga representationer som bör ha betraktas som irrelevant istället för att lära sig form. Modell D lärde sig oföränderliga representationer av form men också irrelevanta representationer. Modellerna F, G och H lyckades lära sig oföränderliga representationer av form med varierande intensitet, samtidigt som de resterande egenskaperna betraktades som irrelevant. Beskärning och färgförvrängande datamodifieringarna gör således att självövervakande modeller kan lära sig oföränderliga representationer av egenskaper relaterade till relevant innehåll. Specifika inställningar för datamodifieringar måste dock vara lämpliga för datasetet.

This degree project proposes the design and the implementation of a novel systemfor deep analysis on large-scale datasets of hand poses. The system consists of a set ofmodules for automatic redundancy removal, classification, statistical analysis andvisualization of large-scale datasets based on their content characteristics. In thisproject, work is performed on the specific use case of images of hand movements infront of smartphone cameras. The characteristics of the images are investigated, andthe images are pre-processed to reduce repetitive content and noise in the data. Twodifferent design paradigms for content analysis and image classification areemployed, a computer vision pipeline and a deep learning pipeline. The computervision pipeline incorporates several stages of image processing including imagesegmentation, hand detection as well as feature extraction followed by a classificationstage. The deep learning pipeline utilizes a convolutional neural network forclassification. For industrial applications with high diversity on data content, deeplearning is suggested for image classification and computer vision is recommendedfor feature analysis. Finally, statistical analysis is performed to visually extractrequired information about hand features and diversity of the classified data. Themain contribution of this work lies in the customization of computer vision and deeplearning tools for the design and the implementation of a hybrid system for deep dataanalysis. / Detta examensprojekt föreslår design och implementering av ett nytt system för djup analys av storskaliga datamängder av handställningar. Systemet består av en uppsättning moduler för automatisk borttagning av redundans, klassificering, statistisk analys och visualisering av storskaliga dataset baserade på deras egenskaper. I det här projektet utförs arbete på det specifika användningsområdet för bilder av handrörelser framför smarttelefonkameror. Egenskaperna hos bilderna undersöks, och bilderna förbehandlas för att minska repetitivt innehåll och ljud i data. Två olika designparadigmer för innehållsanalys och bildklassificering används, en datorvisionspipeline och en djuplärningsrörledning. Datasynsrörledningen innehåller flera steg i bildbehandling, inklusive bildsegmentering, handdetektering samt funktionen extraktion följt av ett klassificeringssteg. Den djupa inlärningsrörledningen använder ett fällningsnätverk för klassificering. För industriella applikationer med stor mångfald på datainnehåll föreslås djupinlärning för bildklassificering och vision rekommenderas för funktionsanalys. Slutligen utförs statistisk analys för att visuellt extrahera nödvändig information om handfunktioner och mångfald av klassificerade data. Huvuddelen av detta arbete ligger i anpassningen av datasyn och djupa inlärningsverktyg för design och implementering av ett hybridsystem för djup dataanalys.