Classroom note taking is a fundamental task performed by learners on a daily basis. These notes provide learners with valuable offline study material, especially in the case of more difficult subjects. The use of class notes has been found to not only provide students with a better learning experience, but also leads to an overall higher academic performance. In a previous study, an increase of 10.5% in student grades was observed after these students had been provided with multimedia class notes. This is not surprising, as other studies have found that the rate of successful transfer of information to humans increases when provided with both visual and audio information. Note taking might seem like an easy task; however, students with hearing impairments, visual impairments, physical impairments, learning disabilities or even non-native listeners find this task very difficult to impossible. It has also been reported that even non-disabled students find note taking time consuming and that it requires a great deal of mental effort while also trying to pay full attention to the lecturer. This is illustrated by a study where it was found that college students were only able to record ~40% of the data presented by the lecturer. It is thus reasonable to expect an automatic way of generating class notes to be beneficial to all learners. Lecture transcription (LT) systems are used in educational environments to assist learners by providing them with real-time in-class transcriptions or recordings and transcriptions for offline use. Such systems have already been successfully implemented in the developed world where all required resources were easily obtained. These systems are typically trained on hundreds to thousands of hours of speech while their language models are trained on millions or even hundreds of millions of words. These amounts of data are generally not available in the developing world. In this dissertation, a number of approaches toward the development of LT systems in resource-scarce environments are investigated. We focus on different approaches to obtaining sufficient amounts of well transcribed data for building acoustic models, using corpora with few transcriptions and of variable quality. One approach investigates the use of alignment using a dynamic programming phone string alignment procedure to harvest as much usable data as possible from approximately transcribed speech data. We find that target-language acoustic models are optimal for this purpose, but encouraging results are also found when using models from another language for alignment. Another approach entails using unsupervised training methods where an initial low accuracy recognizer is used to transcribe a set of untranscribed data. Using this poorly transcribed data, correctly recognized portions are extracted based on a word confidence threshold. The initial system is retrained along with the newly recognized data in order to increase its overall accuracy. The initial acoustic models are trained using as little as 11 minutes of transcribed speech. After several iterations of unsupervised training, a noticeable increase in accuracy was observed (47.79% WER to 33.44% WER). Similar results were however found (35.97% WER) after using a large speaker-independent corpus to train the initial system. Usable LMs were also created using as few as 17955 words from transcribed lectures; however, this resulted in large out-of-vocabulary rates. This problem was solved by means of LM interpolation. LM interpolation was found to be very beneficial in cases where subject specific data (such as lecture slides and books) was available. We also introduce our NWU LT system, which was developed for use in learning environments and was designed using a client/server based architecture. Based on the results found in this study we are confident that usable models for use in LT systems can be developed in resource-scarce environments. / MSc (Computer Science), North-West University, Vaal Triangle Campus, 2014

Classroom note taking is a fundamental task performed by learners on a daily basis. These notes provide learners with valuable offline study material, especially in the case of more difficult subjects. The use of class notes has been found to not only provide students with a better learning experience, but also leads to an overall higher academic performance. In a previous study, an increase of 10.5% in student grades was observed after these students had been provided with multimedia class notes. This is not surprising, as other studies have found that the rate of successful transfer of information to humans increases when provided with both visual and audio information. Note taking might seem like an easy task; however, students with hearing impairments, visual impairments, physical impairments, learning disabilities or even non-native listeners find this task very difficult to impossible. It has also been reported that even non-disabled students find note taking time consuming and that it requires a great deal of mental effort while also trying to pay full attention to the lecturer. This is illustrated by a study where it was found that college students were only able to record ~40% of the data presented by the lecturer. It is thus reasonable to expect an automatic way of generating class notes to be beneficial to all learners. Lecture transcription (LT) systems are used in educational environments to assist learners by providing them with real-time in-class transcriptions or recordings and transcriptions for offline use. Such systems have already been successfully implemented in the developed world where all required resources were easily obtained. These systems are typically trained on hundreds to thousands of hours of speech while their language models are trained on millions or even hundreds of millions of words. These amounts of data are generally not available in the developing world. In this dissertation, a number of approaches toward the development of LT systems in resource-scarce environments are investigated. We focus on different approaches to obtaining sufficient amounts of well transcribed data for building acoustic models, using corpora with few transcriptions and of variable quality. One approach investigates the use of alignment using a dynamic programming phone string alignment procedure to harvest as much usable data as possible from approximately transcribed speech data. We find that target-language acoustic models are optimal for this purpose, but encouraging results are also found when using models from another language for alignment. Another approach entails using unsupervised training methods where an initial low accuracy recognizer is used to transcribe a set of untranscribed data. Using this poorly transcribed data, correctly recognized portions are extracted based on a word confidence threshold. The initial system is retrained along with the newly recognized data in order to increase its overall accuracy. The initial acoustic models are trained using as little as 11 minutes of transcribed speech. After several iterations of unsupervised training, a noticeable increase in accuracy was observed (47.79% WER to 33.44% WER). Similar results were however found (35.97% WER) after using a large speaker-independent corpus to train the initial system. Usable LMs were also created using as few as 17955 words from transcribed lectures; however, this resulted in large out-of-vocabulary rates. This problem was solved by means of LM interpolation. LM interpolation was found to be very beneficial in cases where subject specific data (such as lecture slides and books) was available. We also introduce our NWU LT system, which was developed for use in learning environments and was designed using a client/server based architecture. Based on the results found in this study we are confident that usable models for use in LT systems can be developed in resource-scarce environments. / MSc (Computer Science), North-West University, Vaal Triangle Campus, 2014

As the world has become more connected meetings have moved online. However, since few have access to studio lighting and uses the embedded webcam the video quality can be far from good. Hence, there is an interest in using a software solution to enhance the video quality in real time. This thesis investigates the feasibility to train a machine learning model to automatically enhance the quality of images. The model must learn without using paired images, since it is difficult to capture images with the exact same content but different quality. Furthermore, the model has to process at least 30 images per second which is a common frequency for videos. Therefore, this thesis investigates the possibility to train a model without paired images and whether such a model can be used in real-time. To answer these questions several sizes of the same model was trained. These were evaluated using six different measures during in order to determine if training without paired data is possible. The models image enhancement capabilities and inference speed were investigated followed by attempts at improving the speed. Finally, different combinations of datasets were investigated to test how well the model generalised to new data. The results show that it is possible to train models for image enhancement without paired data. However, to use such a model in real time a graphics card is needed to reach above 30 images per second.

Au long des dernières décennies, des importants avancements ont été réalisés dans le domaine de la reconnaissance de la parole à grand vocabulaire. Un des défis à relever dans le domaine concerne la réduction des coûts de développement nécessaires pour construire un nouveau système ou adapter un système existant à une nouvelle tâche, langue ou dialecte. Les systèmes de reconnaissance de la parole à l’état de l’art sont basés sur les principes de l’apprentissage statistique, utilisant l’information fournie par deux modèles stochastiques, un modèle acoustique (MA) et un modèle de langue (ML). Les méthodes standards utilisées pour construire ces modèles s’appuient sur deux hypothèses de base : les jeux de données d’apprentissage sont suffisamment grands, et les données d’apprentissage correspondent bien à la tâche cible. Il est bien connu qu’une partie importante des coûts de développement est dû à la préparation des corpora qui remplissent ces deux conditions, l’origine principale des coûts étant la transcription manuelle des données audio. De plus, pour certaines applications, notamment la reconnaissance des langues et dialectes dits "peu dotés", la collecte des données est en soi une mission difficile. Cette thèse a pour but d’examiner et de proposer des méthodes visant à réduire le besoin de transcriptions manuelles des données audio pour une tâche donnée. Deux axes de recherche ont été suivis. Dans un premier temps, des méthodes d’apprentissage dits "non-supervisées" sont explorées. Leur point commun est l’utilisation des transcriptions audio obtenues automatiquement à l’aide d’un système de reconnaissance existant. Des méthodes non-supervisées sont explorées pour la construction de trois des principales composantes des systèmes de reconnaissance. D’abord, une nouvelle méthode d’apprentissage non-supervisée des MAs est proposée : l’utilisation de plusieurs hypothèses de décodage (au lieu de la meilleure uniquement) conduit à des gains de performance substantiels par rapport à l’approche standard. L’approche non-supervisée est également étendue à l’estimation des paramètres du réseau de neurones (RN) utilisé pour l’extraction d’attributs acoustiques. Cette approche permet la construction des modèles acoustiques d’une façon totalement non-supervisée et conduit à des résultats compétitifs en comparaison avec des RNs estimés de façon supervisée. Finalement, des méthodes non-supervisées sont explorées pour l’estimation des MLs à repli (back-off ) standards et MLs neuronaux. Il est montré que l’apprentissage non-supervisée des MLs conduit à des gains de performance additifs (bien que petits) à ceux obtenus par l’apprentissage non-supervisée des MAs. Dans un deuxième temps, cette thèse propose l’utilisation de l’interpolation de modèles comme une alternative rapide et flexible pour la construction des MAs pour une tâche cible. Les modèles obtenus à partir d’interpolation se montrent plus performants que les modèles de base, notamment ceux estimés à échantillons regroupés ou ceux adaptés à la tâche cible. Il est montré que l’interpolation de modèles est particulièrement utile pour la reconnaissance des dialectes peu dotés. Quand la quantité de données d’apprentissage acoustiques du dialecte ciblé est petite (2 à 3 heures) ou même nulle, l’interpolation des modèles conduit à des gains de performances considérables par rapport aux méthodes standards. / One of the outstanding challenges in large vocabulary automatic speech recognition (ASR) is the reduction of development costs required to build a new recognition system or adapt an existing one to a new task, language or dialect. The state-of-the-art ASR systems are based on the principles of the statistical learning paradigm, using information provided by two stochastic models, an acoustic (AM) and a language (LM) model. The standard methods used to estimate the parameters of such models are founded on two main assumptions : the training data sets are large enough, and the training data match well the target task. It is well-known that a great part of system development costs is due to the construction of corpora that fulfill these requirements. In particular, manually transcribing the audio data is the most expensive and time-consuming endeavor. For some applications, such as the recognition of low resourced languages or dialects, finding and collecting data is also a hard (and expensive) task. As a means to lower the cost required for ASR system development, this thesis proposes and studies methods that aim to alleviate the need for manually transcribing audio data for a given target task. Two axes of research are explored. First, unsupervised training methods are explored in order to build three of the main components of ASR systems : the acoustic model, the multi-layer perceptron (MLP) used to extract acoustic features and the language model. The unsupervised training methods aim to estimate the model parameters using a large amount of automatically (and inaccurately) transcribed audio data, obtained thanks to an existing recognition system. A novel method for unsupervised AM training that copes well with the automatic audio transcripts is proposed : the use of multiple recognition hypotheses (rather than the best one) leads to consistent gains in performance over the standard approach. Unsupervised MLP training is proposed as an alternative to build efficient acoustic models in a fully unsupervised way. Compared to cross-lingual MLPs trained in a supervised manner, the unsupervised MLP leads to competitive performance levels even if trained on only about half of the data amount. Unsupervised LM training approaches are proposed to estimate standard back-off n-gram and neural network language models. It is shown that unsupervised LM training leads to additive gains in performance on top of unsupervised AM training. Second, this thesis proposes the use of model interpolation as a rapid and flexible way to build task specific acoustic models. In reported experiments, models obtained via interpolation outperform the baseline pooled models and equivalent maximum a posteriori (MAP) adapted models. Interpolation proves to be especially useful for low resourced dialect ASR. When only a few (2 to 3 hours) or no acoustic data truly matching the target dialect are available for AM training, model interpolation leads to substantial performance gains compared to the standard training methods.

A prática regular de exercícios físicos supervisionados tem sido recomendada devido a seus benefícios na saúde. Porém, a supervisão limita o número de praticantes. Uma alternativa para a promoção da saúde pública é o treinamento de caminhada prescrita de forma individualizada e realizada sem supervisão, mas seus efeitos foram pouco estudados em uma situação real. Além disso, a possível influência do nível inicial de atividade física (AF) e de aptidão física (ApF) sobre os efeitos desse treinamento não é conhecida. Assim, este estudo investigou na situação real do Projeto Exercício e Coração: 1) a relação do nível de AF e de ApF com o risco cardiovascular avaliado de forma isolada e global; 2) o efeito do treinamento de caminhada prescrita de forma individualizada e executada sem supervisão sobre esse risco cardiovascular; e 3) a influência do nível inicial de AF e de ApF nas respostas ao treinamento. O risco cardiovascular isolado foi avaliado pela medida do índice de massa corporal (IMC), circunferência da cintura (CC), glicemia, colesterol total e pressão arterial (PA) sistólica e diastólica, enquanto que o risco global foi calculado pelo escore Z (EZ, somatório do escore z de todos os fatores isolados). O nível de AF foi avaliado pelos minutos semanais de AF de lazer e a ApF pelo resultado do teste de marcha estacionária dividido em quartis (Q1 = pior ApF e Q4 = melhor ApF). O IMC e a CC foram menores no grupo muito ativo (MA, +300 min/sem AF) do que no inativo (I, sem nenhuma AF de lazer), enquanto que o EZ foi menor no grupo MA que no I e no ativo (A, 150 a 299 min/sem de AF). Além disso, o IMC e a glicemia foram menores no Q4 que no Q1, a CC foi menor no Q2, Q3 e Q4 que no Q1, e o EZ foi menor no Q3 e Q4 que no Q1 e no Q4 que no Q2. O treinamento de caminhada diminuiu significantemente o IMC, CC, PA sistólica e EZ na amostra total. Além disso, ele reduziu significantemente todos os indicadores de risco específicos em subamostras com valores alterados, com exceção da glicemia. Para finalizar, o efeito do treinamento de caminhada no risco cardiovascular isolado e global foi semelhante nos indivíduos MA e I e nos indivíduos do Q1 e Q4 de ApF. Assim, é possível concluir que: 1) em participantes de uma situação real de promoção de AF para a saúde há associação inversa entre os níveis de AF e de ApF com marcadores de obesidade, glicemia (só ApF) e com o risco cardiovascular global; 2) o treinamento de caminhada prescrita de forma individual e executada sem supervisão em uma situação real reduz alguns fatores de risco isolados, principalmente quando eles estão alterados, e diminui o risco cardiovascular global; e 3) nem o nível inicial de AF nem o de ApF afetam os efeitos de um treinamento de caminhada prescrito de forma individualizada e executado sem supervisão em uma situação real sobre o risco cardiovascular / The regular practice of supervised physical exercise has been recommended due to its benefits on health. However, supervision limits the number of practitioners. An interesting alternative for the promotion of public health is the walking training prescribed individually and executed without supervision, however its effects have been poorly investigated under real situations. In addition, the possible influence of the initial level of physical activity (PA) and physical fitness (PF) on these effects are unknown. Thus, this study investigated under a real situation of the \"Exercise and Heart Project\": 1) the relationship between level of PA and PF with cardiovascular risk evaluated by isolated factors and globally; 2) the effects of walking training prescribed individually and executed without supervision on cardiovascular risk; and 3) the influence of initial level of PA and PF on the responses to walking training. Isolated cardiovascular risk was analyzed by the measurement of body mass index (BMI), waist circumference (WC), blood glucose, total cholesterol, and systolic and diastolic blood pressure (BP), while global risk was calculated by Z score (ZS, sum of Z score of all the factors). PA level was evaluated by weekly minutes of leisure time PA, and PF was evaluated by the results in the 2 minutes step test divided in quartiles (Q1 being the lowest PF and Q4 the highest PF). BMI and WC were significantly lower in the very active (VA, +300 min/week of PA) than in the inactive group (I, no leisure time PA), while ZS was lower in the VA than in the I and active groups (A, between 150 and 299 min/week of PA). In addition, BMI and blood glucose was significantly lower in Q4 than Q1, WC was lower in Q2, Q3 and Q4 than in Q1, and ZS was lower in Q3 and Q4 than Q1 and in Q4 than Q2. Walking training significantly decreased BMI, WC, systolic BP and ZS in the total sample. Also, it decreased all specific risk factors, with exception of blood glucose in subgroups with altered values. Finally, the effects of walking training on isolated risk factors and on ZS were similar in VA and I groups as well as in Q1 and Q4 groups. Thus, it is possible to conclude that: 1) in participants of a real intervention for health promotion, there is an inverse association between the PA and PF levels with the obesity markers, blood glucose (only PF) and global cardiovascular risk; 2) the walking training prescribed individually and executed without supervision in a real situation reduces some isolated cardiovascular risk factors, especially when they are altered, and decreases global cardiovascular risk; and 3) neither the initial level of PA nor the initial levels of PF affects the effects of walking training prescribed individually and executed without supervision in a real situation on cardiovascular risk

This thesis describes new gene finding methods for eukaryotic gene prediction. The current methods for deriving model parameters for gene prediction algorithms are based on curated or experimentally validated set of genes or gene elements. These training sets often require time and additional expert efforts especially for the species that are in the initial stages of genome sequencing. Unsupervised training allows determination of model parameters from anonymous genomic sequence with. The importance and the practical applicability of the unsupervised training is critical for ever growing rate of eukaryotic genome sequencing. Three distinct training procedures are developed for diverse group of eukaryotic species. GeneMark-ES is developed for species with strong donor and acceptor site signals such as Arabidopsis thaliana, Caenorhabditis elegans and Drosophila melanogaster. The second version of the algorithm, GeneMark-ES-2, introduces enhanced intron model to better describe the gene structure of fungal species with posses with relatively weak donor and acceptor splice sites and well conserved branch point signal. GeneMark-LE, semi-supervised training approach is designed for eukaryotic species with small number of introns. The results indicate that the developed unsupervised training methods perform well as compared to other training methods and as estimated from the set of genes supported by EST-to-genome alignments. Analysis of novel genomes reveals interesting biological findings and show that several candidates of under-annotated and over-annotated fungal species are present in the current set of annotated of fungal genomes.

La segmentation et regroupement en locuteurs (SRL) de collection cherche à répondre à la question « qui parle quand ? » dans une collection de documents multimédia. C’est un prérequis indispensable à l’indexation des contenus audiovisuels. La tâche de SRL consiste d’abord à segmenter chaque document en locuteurs, avant de les regrouper à l'échelle de la collection. Le but est de positionner des labels anonymes identifiant les locuteurs, y compris ceux apparaissant dans plusieurs documents, sans connaître à l'avance ni leur identité ni leur nombre. La difficulté posée par le regroupement en locuteurs à l'échelle d'une collection est le problème de la variabilité intra-locuteur/inter-document : selon les documents, un locuteur peut parler dans des environnements acoustiques variés (en studio, dans la rue...). Cette thèse propose deux méthodes pour pallier le problème. D'une part, une nouvelle méthode de compensation neuronale de variabilité est proposée, utilisant le paradigme de triplet-loss pour son apprentissage. D’autre part, un procédé itératif d'adaptation non supervisée au domaine est présenté, exploitant l'information, même imparfaite, que le système acquiert en traitant des données, pour améliorer ses performances sur le domaine acoustique cible. De plus, de nouvelles méthodes d'analyse en locuteurs des résultats de SRL sont étudiées, pour comprendre le fonctionnement réel des systèmes, au-delà du classique taux d'erreur de SRL (Diarization Error Rate ou DER). Les systèmes et méthodes sont évalués sur deux émissions télévisées d'une quarantaine d'épisodes, pour les architectures de SRL globale ou incrémentale, à l'aide de la modélisation locuteur à l'état de l'art. / The task of speaker diarization and linking aims at answering the question "who speaks and when?" in a collection of multimedia recordings. It is an essential step to index audiovisual contents. The task of speaker diarization and linking firstly consists in segmenting each recording in terms of speakers, before linking them across the collection. Aim is, to identify each speaker with a unique anonymous label, even for speakers appearing in multiple recordings, without any knowledge of their identity or number. The challenge of the cross-recording linking is the modeling of the within-speaker/across-recording variability: depending on the recording, a same speaker can appear in multiple acoustic conditions (in a studio, in the street...). The thesis proposes two methods to overcome this issue. Firstly, a novel neural variability compensation method is proposed, using the triplet-loss paradigm for training. Secondly, an iterative unsupervised domain adaptation process is presented, in which the system exploits the information (even inaccurate) about the data it processes, to enhance its performances on the target acoustic domain. Moreover, novel ways of analyzing the results in terms of speaker are explored, to understand the actual performance of a diarization and linking system, beyond the well-known Diarization Error Rate (DER). Systems and methods are evaluated on two TV shows of about 40 episodes, using either a global, or longitudinal linking architecture, and state of the art speaker modeling (i-vector).

A prática regular de exercícios físicos supervisionados tem sido recomendada devido a seus benefícios na saúde. Porém, a supervisão limita o número de praticantes. Uma alternativa para a promoção da saúde pública é o treinamento de caminhada prescrita de forma individualizada e realizada sem supervisão, mas seus efeitos foram pouco estudados em uma situação real. Além disso, a possível influência do nível inicial de atividade física (AF) e de aptidão física (ApF) sobre os efeitos desse treinamento não é conhecida. Assim, este estudo investigou na situação real do Projeto Exercício e Coração: 1) a relação do nível de AF e de ApF com o risco cardiovascular avaliado de forma isolada e global; 2) o efeito do treinamento de caminhada prescrita de forma individualizada e executada sem supervisão sobre esse risco cardiovascular; e 3) a influência do nível inicial de AF e de ApF nas respostas ao treinamento. O risco cardiovascular isolado foi avaliado pela medida do índice de massa corporal (IMC), circunferência da cintura (CC), glicemia, colesterol total e pressão arterial (PA) sistólica e diastólica, enquanto que o risco global foi calculado pelo escore Z (EZ, somatório do escore z de todos os fatores isolados). O nível de AF foi avaliado pelos minutos semanais de AF de lazer e a ApF pelo resultado do teste de marcha estacionária dividido em quartis (Q1 = pior ApF e Q4 = melhor ApF). O IMC e a CC foram menores no grupo muito ativo (MA, +300 min/sem AF) do que no inativo (I, sem nenhuma AF de lazer), enquanto que o EZ foi menor no grupo MA que no I e no ativo (A, 150 a 299 min/sem de AF). Além disso, o IMC e a glicemia foram menores no Q4 que no Q1, a CC foi menor no Q2, Q3 e Q4 que no Q1, e o EZ foi menor no Q3 e Q4 que no Q1 e no Q4 que no Q2. O treinamento de caminhada diminuiu significantemente o IMC, CC, PA sistólica e EZ na amostra total. Além disso, ele reduziu significantemente todos os indicadores de risco específicos em subamostras com valores alterados, com exceção da glicemia. Para finalizar, o efeito do treinamento de caminhada no risco cardiovascular isolado e global foi semelhante nos indivíduos MA e I e nos indivíduos do Q1 e Q4 de ApF. Assim, é possível concluir que: 1) em participantes de uma situação real de promoção de AF para a saúde há associação inversa entre os níveis de AF e de ApF com marcadores de obesidade, glicemia (só ApF) e com o risco cardiovascular global; 2) o treinamento de caminhada prescrita de forma individual e executada sem supervisão em uma situação real reduz alguns fatores de risco isolados, principalmente quando eles estão alterados, e diminui o risco cardiovascular global; e 3) nem o nível inicial de AF nem o de ApF afetam os efeitos de um treinamento de caminhada prescrito de forma individualizada e executado sem supervisão em uma situação real sobre o risco cardiovascular / The regular practice of supervised physical exercise has been recommended due to its benefits on health. However, supervision limits the number of practitioners. An interesting alternative for the promotion of public health is the walking training prescribed individually and executed without supervision, however its effects have been poorly investigated under real situations. In addition, the possible influence of the initial level of physical activity (PA) and physical fitness (PF) on these effects are unknown. Thus, this study investigated under a real situation of the \"Exercise and Heart Project\": 1) the relationship between level of PA and PF with cardiovascular risk evaluated by isolated factors and globally; 2) the effects of walking training prescribed individually and executed without supervision on cardiovascular risk; and 3) the influence of initial level of PA and PF on the responses to walking training. Isolated cardiovascular risk was analyzed by the measurement of body mass index (BMI), waist circumference (WC), blood glucose, total cholesterol, and systolic and diastolic blood pressure (BP), while global risk was calculated by Z score (ZS, sum of Z score of all the factors). PA level was evaluated by weekly minutes of leisure time PA, and PF was evaluated by the results in the 2 minutes step test divided in quartiles (Q1 being the lowest PF and Q4 the highest PF). BMI and WC were significantly lower in the very active (VA, +300 min/week of PA) than in the inactive group (I, no leisure time PA), while ZS was lower in the VA than in the I and active groups (A, between 150 and 299 min/week of PA). In addition, BMI and blood glucose was significantly lower in Q4 than Q1, WC was lower in Q2, Q3 and Q4 than in Q1, and ZS was lower in Q3 and Q4 than Q1 and in Q4 than Q2. Walking training significantly decreased BMI, WC, systolic BP and ZS in the total sample. Also, it decreased all specific risk factors, with exception of blood glucose in subgroups with altered values. Finally, the effects of walking training on isolated risk factors and on ZS were similar in VA and I groups as well as in Q1 and Q4 groups. Thus, it is possible to conclude that: 1) in participants of a real intervention for health promotion, there is an inverse association between the PA and PF levels with the obesity markers, blood glucose (only PF) and global cardiovascular risk; 2) the walking training prescribed individually and executed without supervision in a real situation reduces some isolated cardiovascular risk factors, especially when they are altered, and decreases global cardiovascular risk; and 3) neither the initial level of PA nor the initial levels of PF affects the effects of walking training prescribed individually and executed without supervision in a real situation on cardiovascular risk

Machine learning has seen a rapid progress the last couple of decades, with more and more powerful neural network models continuously being presented. These neural networks require large amounts of data to train them. Labelled data is especially in great demand, but due to the time consuming and costly nature of data labelling, there exists a scarcity for labelled data, whereas there usually is an abundance of unlabelled data. In some cases, data from a certain distribution, or domain, is labelled, whereas the data we actually want to optimise our model on is unlabelled and from another domain. This falls under the umbrella of domain adaptation and the purpose of this thesis is to train a network using domain-adversarial training on eye image datasets consisting of a labelled source domain and an unlabelled target domain, with the goal of performing well on target data, i.e., overcoming the domain gap. This was done on two different datasets: a proprietary dataset from Tobii with real images and the public U2Eyes dataset with synthetic data. When comparing domain-adversarial training to a baseline model trained conventionally on source data and a oracle model trained conventionally on target data, the proposed DAT-ResNet model outperformed the baseline on both datasets. For the Tobii dataset, DAT-ResNet improved the Huber loss by 22.9% and the Intersection over Union (IoU) by 7.6%, and for the U2Eyes dataset, DAT-ResNet improved the Huber loss by 67.4% and the IoU by 37.6%. Furthermore, the IoU measures were extended to also include the portion of predicted ellipsis with no intersection with the corresponding ground truth ellipsis – referred to as zero-IoUs. By this metric, the proposed model improves the percentage of zero-IoUs by 34.9% on the Tobii dataset and by 90.7% on the U2Eyes dataset. / Maskininlärning har sett en snabb utveckling de senaste decennierna med mer och mer kraftfulla neurala nätverk-modeller presenterades kontinuerligt. Dessa neurala nätverk kräver stora mängder data för att tränas. Data med etiketter är det framförallt stor efterfrågan på, men på grund av det är tidskrävande och kostsamt att etikettera data så finns det en brist på sådan data medan det ofta finns ett överflöd av data utan etiketter. I vissa fall så är data från en viss fördelning, eller domän, etiketterad, medan datan som vi faktiskt vill optimera vår modell efter saknar etiketter och är från en annan domän. Det här faller under området domänadaptering och målet med det här arbetet är att träna ett nätverk genom att använda domänmoststående träning på dataset med ögonbilder som har en källdomän med etiketter och en måldomän utan etiketter, där målet är att prestera bra på data från måldomänen, i.e., att lösa ett domänadapteringsproblem. Det här gjordes på två olika dataset: ett dataset som ägs av Tobii med riktiga ögonbilder och det offentliga datasetet U2Eyes med syntetiska bilder. När domänadapteringsmodellen jämförs med en basmodell tränad konventionellt på källdata och en orakelmodell tränad konventionellt på måldata, så utklassar den presenterade DAT-ResNet-modellen basmodellen på båda dataseten. På Tobii-datasetet så förbättrade DAT-ResNet förlusten med 22.9% och Intersection over Union (IoU):n med 7.6%, och på U2Eyes-datasetet, förbättrade DAT-ResNet förlusten med 67.4% och IoU:n med 37.6%. Dessutom så utökades IoU-måtten till att också innefatta andelen av förutspådda ellipser utan något överlapp med tillhörande grundsanningsellipser – refererat till som noll-IoU:er. Enligt detta mått så förbättrar den föreslagna modellen noll-IoU:erna med 34.9% på Tobii-datasetet och 90.7% på U2Eyes-datasetet.