Dissertations / Theses on the topic 'Surface learning'

Cette thèse se concentre sur le développement demodèles de représentation dense d’objets 3-D àpartir d’images. L’objectif de ce travail estd’améliorer les modèles surfaciques 3-D fournispar les systèmes de vision par ordinateur, enutilisant de nouveaux éléments tirés des images,plutôt que les annotations habituellementutilisées, ou que les modèles basés sur unedivision de l’objet en différents parties.Des réseaux neuronaux convolutifs (CNNs) sontutilisés pour associer de manière dense les pixelsd’une image avec les coordonnées 3-D d’unmodèle de l’objet considéré. Cette méthodepermet de résoudre très simplement unemultitude de tâches de vision par ordinateur,telles que le transfert d’apparence, la localisationde repères ou la segmentation sémantique, enutilisant la correspondance entre une solution surle modèle surfacique 3-D et l’image 2-Dconsidérée. On démontre qu’une correspondancegéométrique entre un modèle 3-D et une imagepeut être établie pour le visage et le corpshumains
This thesis addresses the task of establishing adense correspondence between an image and a 3Dobject template. We aim to bring vision systemscloser to a surface-based 3D understanding ofobjects by extracting information that iscomplementary to existing landmark- or partbasedrepresentations.We use convolutional neural networks (CNNs)to densely associate pixels with intrinsiccoordinates of 3D object templates. Through theestablished correspondences we effortlesslysolve a multitude of visual tasks, such asappearance transfer, landmark localization andsemantic segmentation by transferring solutionsfrom the template to an image. We show thatgeometric correspondence between an imageand a 3D model can be effectively inferred forboth the human face and the human body

In traditional industrial manufacturing, due to the limitations of science and technology, manual inspection methods are still used to detect product surface defects. This method is slow and inefficient due to manual limitations and backward technology. The aim of this thesis is to research whether it is possible to automate this using modern computer hardware and image classification of defects using different deep learning methods. The report concludes, based on results from controlled experiments, that it is possible to achieve a dice coefficient of more than 81%.

This research aims to investigate an assertion, endorsed by a range of commentators, that multimedia teaching and learning approaches encourage learners to adopt a richer, creative and deeper level of understanding and participation within the learning environment than traditional teaching and learning methods. The thesis examines this assertion by investigating one type of multimedia activity defined (for the purposes of this research) as a digital audio video assessment (DAVA). Data was collected using a constructivist epistemology, interpretative and naturalistic perspective using primarily a qualitative methodology. Three types of data collection methods were used to collect data from thirteen Diploma of Event Management students from William Angliss TAFE. Firstly, participants completed the Biggs Study Process Questionnaire (2001) which is a predictor of deep and surface learning preference. Each participant then engaged in a semi-structured interview that elicited participant's self-declared learning preferences and their approaches to completion of the DAVA. These data sources were then compared. Six factors that are critical in informing the way that the participants approached the DAVA emerged from the analysis of the data. Based on these findings it is concluded that the DAVA does not restrict, inhibit or negatively influence a participants learning preference. Learners with a pre-existing, stable learning preference are likely to adopt a learning approach that is consisten t with their preference. Participants that have a learning preference that is less stable (more flexible) may adopt either a surface or deep approach depending on the specific task, activity or assessment.

The release of crude oil or other petroleum based products into marine habitats can have a devastating impact on the environment as well as the local economies that rely on these waters for commercial fishing and tourism. The Deepwater Horizon catastrophe that started on April 20th 2010 leaked an estimated 4.4 million barrels of crude oil into the Gulf of Mexico over a 3 month period threatening thousands of species and crippling the gulf coast. The National Oceanic and Atmospheric Administration (NOAA) used several satellite remote sensing technologies to manually track and predict the extent and location of oil on the surface of the gulf waters. This thesis proposes a methodology to automatically identify surface oil using an unsupervised clustering algorithm an compares the discovered regions of oil to the reports generated by NOAA during the incident. The fuzzy c-means clustering algorithm is used to partition the satellite image pixels into groups that represent either oil or not oil. A variety of MODIS data features and image analyzing techniques have been explored to produce the most accurate set of regions.

The task of automatically segmenting 3-D surfaces representing object boundaries is important in quantitative analysis of volumetric images, which plays a vital role in numerous biomedical applications. For the diagnosis and management of disease, segmentation of images of organs and tissues is a crucial step for the quantification of medical images. Segmentation finds the boundaries or, limited to the 3-D case, the surfaces, that separate regions, tissues or areas of an image, and it is essential that these boundaries approximate the true boundary, typically by human experts, as closely as possible. Recently, graph-based methods with a global optimization property have been studied and used for various applications. Sepecifically, the state-of-the-art graph search (optimal surface segmentation) method has been successfully used for various such biomedical applications. Despite their widespread use for image segmentation, real world medical image segmentation problems often pose difficult challenges, wherein graph based segmentation methods in its purest form may not be able to perform the segmentation task successfully. This doctoral work has a twofold objective. 1)To identify medical image segmentation problems which are difficult to solve using existing graph based method and develop novel methods by employing graph search as a building block to improve segmentation accuracy and efficiency. 2) To develop a novel multiple surface segmentation strategy using deep learning which is more computationally efficient and generic than the exisiting graph based methods, while eliminating the need for human expert intervention as required in the current surface segmentation methods. This developed method is possibly the first of its kind where the method does not require and human expert designed operations. To accomplish the objectives of this thesis work, a comprehensive framework of graph based and deep learning methods is proposed to achieve the goal by successfully fulfilling the follwoing three aims. First, an efficient, automated and accurate graph based method is developed to segment surfaces which have steep change in surface profiles and abrupt distance changes between two adjacent surfaces. The developed method is applied and validated on intra-retinal layer segmentation of Spectral Domain Optical Coherence Tomograph (SD-OCT) images of eye with Glaucoma, Age Related Macular Degneration and Pigment Epithelium Detachment. Second, a globally optimal graph based method is developed to attain subvoxel and super resolution accuracy for multiple surface segmentation problem while imposing convex constraints. The developed method was applied to layer segmentation of SD-OCT images of normal eye and vessel walls in Intravascular Ultrasound (IVUS) images. Third, a deep learning based multiple surface segmentation is developed which is more generic, computaionally effieient and eliminates the requirement of human expert interventions (like transformation designs, feature extrraction, parameter tuning, constraint modelling etc.) required by existing surface segmentation methods in varying capacities. The developed method was applied to SD-OCT images of normal and diseased eyes, to validate the superior segmentaion performance, computation efficieny and the generic nature of the framework, compared to the state-of-the-art graph search method.

Analysis of surface models reconstructed from human MR images gives re- searchers the ability to quantify the shape and size of the cerebral cortex. Increasing the reliability of automatic reconstructions would increase the precision and, therefore, power of studies utilizing cortical surface models. We looked at four different workflows for reconstructing cortical surfaces: 1) BAW + LOGIMSOS- B; 2) FreeSurfer + LOGISMOS-B; 3) BAW + FreeSurfer + Machine Learning + LOGISMOS-B; 4) Standard FreeSurfer(Dale et al. 1999). Workflows 1-3 were developed in this project. Workflow 1 utilized both BRAINSAutoWorkup(BAW)(Kim et al. 2015) and a surface reconstruction tool called LOGISMOS-B(Oguz et al. 2014). Workflow 2 added LOGISMOS-B to a custom built FreeSurfer workflow that was highly optimized for parallel processing. Workflow 3 combined workflows 1 and 2 and added random forest classifiers for predicting the edges of the cerebral cortex. These predictions were then fed into LOGISMOS-B as the cost function for graph segmentation. To compare these work- flows, a dataset of 578 simulated cortical volume changes was created from 20 different sets of MR scans. The workflow utilizing machine learning (workflow 3) produced cortical volume changes with the least amount of error when compared to the known volume changes from the simulations. Machine learning can be effectively used to help reconstruct cortical surfaces that more precisely track changes in the cerebral cortex. This research could be used to increase the power of future projects studying correlations between cortical morphometrics and neurological health.

When an engineering student attends four or five years of college to become a professional engineer one makes the assumption that they approach this learning process in such a way to gain the most knowledge possible. The purpose of this study is to measure the learning approach (deep versus surface) of first-year engineering students, test the impact of two interventions (journaling and learning strategy awareness) on increasing the deep approach to learning, and determine the relationship of the approach to learning on retention within an engineering program. The study was conducted using a quantitative self-reporting instrument to measure surface and deep learning at the beginning and end of the first and second semesters of the freshman year in an engineering program. Retention was measured as the continuous enrollment of a student in the second semester of the first-year engineering program. Results indicate that the first-year engineering students have a slightly higher level of the deep approach to learning than a surface approach to learning when they begin college. However, the results also indicate that the deep approach to learning decreased during the first semester and during the second semester of their freshman year. A student's approach to learning can be impacted by their prior knowledge, the teaching context, the institutional context or the motivation of the student. Results surrounding the learning strategies intervention also indicate that the first-year engineering students do not possess the strong learning strategies that are anticipated from students accepted into an engineering program with stringent application requirements. Finally, results indicate that a deep approach to learning appears to have a positive relationship and a surface approach to learning appears to have a negative relationship to retention in an engineering program. This study illustrates that incorporating learning theory and the use of current learning strategy measurements contributes to the understanding of a freshman engineering student's approach to learning. The understanding of the engineering student's approach to learning benefits faculty in establishing curriculum and pedagogical design. The benefit to the student is in understanding more about themselves as a learner.

Abstract Training a visual inspection device is not straightforward but suffers from the high variation in material to be inspected. This variation causes major difficulties for a human, and this is directly reflected in classifier training. Many inspection devices utilize rule-based classifiers the building and training of which rely mainly on human expertise. While designing such a classifier, a human tries to find the questions that would provide proper categorization. In training, an operator tunes the classifier parameters, aiming to achieve as good classification accuracy as possible. Such classifiers require lot of time and expertise before they can be fully utilized. Supervised classifiers form another common category. These learn automatically from training material, but rely on labels that a human has set for it. However, these labels tend to be inconsistent and thus reduce the classification accuracy achieved. Furthermore, as class boundaries are learnt from training samples, they cannot in practise be later adjusted if needed. In this thesis, a visual based training method is presented. It avoids the problems related to traditional training methods by combining a classifier and a user interface. The method relies on unsupervised projection and provides an intuitive way to directly set and tune the class boundaries of high-dimensional data. As the method groups the data only by the similarities of its features, it is not affected by erroneous and inconsistent labelling made for training samples. Furthermore, it does not require knowledge of the internal structure of the classifier or iterative parameter tuning, where a combination of parameter values leading to the desired class boundaries are sought. On the contrary, the class boundaries can be set directly, changing the classification parameters. The time need to take such a classifier into use is small and tuning the class boundaries can happen even on-line, if needed. The proposed method is tested with various experiments in this thesis. Different projection methods are evaluated from the point of view of visual based training. The method is further evaluated using a self-organizing map (SOM) as the projection method and wood as the test material. Parameters such as accuracy, map size, and speed are measured and discussed, and overall the method is found to be an advantageous training and classification scheme.

Understanding road surface conditions is an important component in active vehicle safety. Estimations can be achieved through image classification using increasingly popular convolutional neural networks (CNNs). In this paper, we explore the effects of multi-task learning by creating CNNs capable of simultaneously performing the two tasks road surface condition classification (RSCC) and road scene semantic segmentation (RSSS). A multi-task network, containing a shared feature extractor (VGG16, ResNet-18, ResNet-101) and two taskspecific network branches, is built and trained using the Road-Conditions and Cityscapes datasets. We reveal that utilizing task-dependent homoscedastic uncertainty in the learning process improvesmulti-task model performance on both tasks. When performing task adaptation, using a small set of additional data labeled with semantic information, we gain considerable RSCC improvements on complex models. Furthermore, we demonstrate increased model generalizability in multi-task models, with up to 12% higher F1-score compared to single-task models.

Belt conveyor accidents are mainly attributed to negligence of safety procedures during maintenance work. Entanglement, falling from heights, & collapse of structure or loose materials are the main cause of accidents. While performing maintenance tasks such as cleaning, installation and repair, belt alignment and so on (Lucas et. al. 2007).

Current industry safety programs provide general guidelines for safety training, but do not require any specific training program structure (Shultz, 2002 and Shultz, 2003). For example MSHA (Mine Safety and Health Administration) only requires 24 hours of training. Typically this training is broken down into four hours of training before the employee starts work, the remaining 20 hours has to be performed within the first sixty days of work (Goldbeck, 2003). The information collected through site visits showed that in addition to completing MSHA safety training requirements companies try to reinforce safety issues through daily and weekly safety meetings on job sites. Due to lack of a required safety training structure, every company is independent in terms of their training format that they follow to train their new and experienced work force. As a result, safety engineers depend heavily on in-house safety programs (e.g. audio-video presentations) to deliver the required training hours specified by MSHA for miners.

Based on a review of current training methods this research identifies four problems; existing training methods to educate miners about dangers involved in conveyor belt environments are mainly passive, safety related information in scattered in various media such as images, videos, paper manuals, etc., access to information in current format is difficult, and updating information is difficult.

This research addressed these identified problems by devising a new approach of learning to augment existing methods of training and evaluate the potential of this concept as a safety-training tool. Research has shown that individuals have their own learning style in which they can increase their retention and stimulate their cognitive learning. The proposed work addresses issues relative to passive vs. active learning and classroom-based vs. self-paced training by developing and implementing an interactive multimedia-based safety-training tool called the Digital Safety Manual (DSM). After the DSM was developed it was put through a series of usability evaluation and subjective analysis to measure the potential of the concept. The evaluation and subjective analysis involved both the novice and expert users.

The results that were yield after the evaluations and subjective analysis shows that the DSM has more learning advantages than the typical training methods and it can be used as a supplementary training method to complement the current approaches of training.
Master of Science

This thesis addresses the use of surface electrode arrays to regulate the stimulation applied to the hand and wrist muscles in order to induce hand movement to desired posture. Electrode array-based electric stimulation is a relatively novel and promising rehabilitation technology, due to its potential to deliver selective stimulation signal to underlying muscles via chosen elements of the arrays. A general control strategy developed in this thesis embeds optimisation methods for selection of appropriate elements of the electrode array with iterative learning control. In iterative learning control, the patient makes repeated attempts to complete a predefined task with the aim of gradually decreasing the error between the movement performed and desired one. A number of different gradient-based methods, such as penalty method and sparse optimisation methods has been developed based on theoretical and experimental findings. These methods are used to find a sparse input vector, which is employed to select only those array elements that are critical to task completion within iterative learning control framework. Experimental results using multi-channel stimulation and 40 element surface electrode array confirm accurate tracking of selected hand postures. Based on the experimental results and the existing literature, a new system for the hand and wrist restoration has been designed. The key element of the system is a game-based task oriented training environment designed for a wide group of patients, including patients with spasticity and hemiplegia.

We develop a theoretical framework for a priori estimation of catalytic activity of metal nanoparticles using geometry-based reactivity descriptors of surface atoms and kinetic analysis of reaction pathways at various types of active sites. We show that orbitalwise coordination numbers 𝐶𝑁^α (α = 𝑠 or 𝑑) can be used to predict chemical reactivity of a metal site (e.g., adsorption energies of critical reaction intermediates) by being aware of the neighboring chemical environment, outperforming their regular (𝐶𝑁) and generalized (𝐶̅𝑁̅) counterparts with little added computational cost. Here we include two examples to illustrate this method: CO oxidation on Au (5𝑑¹⁰6𝑠¹) and O₂ reduction on Pt (5𝑑⁹6𝑠¹). We also employ Bayesian learning and the Newns-Anderson model to advance the fundamental understanding of adsorbate-surface interactions on metal nanocatalysts, paving the path toward adsorbate-specific tuning of catalysis.
Doctor of Philosophy
The interactions between reaction intermediates and catalysts should be neither too strong nor too weak for catalytic optimization. This Sabatiers principle arising from the scaling relations among the energetics of reacting species at geometrically similar sites, provides the conceptual basis for designing improved catalysts, but imposes volcano-type limitations on the attainable catalytic activity and selectivity. One of the greatest challenges faced by the catalysis community today is how to develop design strategies and ultimately predictive models of catalytic systems that could circumvent energy scaling relations. This work brings the quantum-chemical modeling and machine learning technique together and develops a novel stochastic modeling approach to rationally design the catalysts with desired properties and bridges our knowledge gap between the empirical kinetics and atomistic mechanisms of catalytic reactions.

Les environnements multi-surfaces (MSE) combinent plusieurs surfaces dans une variété d'arrangements physiques pour former un espace d'information continu. Les grandes surfaces, telles que les tables et les écrans muraux, sont souvent utilisées comme espace partagé pour coordonner les efforts, et les périphériques portables, tels que les tablettes et les smartphones sont considérés comme des espaces personnels prenant en charge les tâches individuelles. Les MSE ont montré des avantages pour soutenir les activités co-implémentées, en particulier celles impliquant une exploration de données riches, telles que des activités collaboratives complexes de résolution de problèmes et de prise de décision. Cependant, la diversité des MSE soulève également des questions et des défis, car différents facteurs de configuration et dispositifs de MSE peuvent être adaptés à différents types d'activités, et le développement d'activités de collaboration dans les MSE reste complexe. Cette dissertation étudie comment les MSE peuvent soutenir la collaboration des utilisateurs en général et l'apprentissage collaboratif en particulier. Elle fournit des informations sur la façon dont la configuration et les facteurs de forme des périphériques des MSE façonnent les comportements collaboratifs des utilisateurs et propose des implications pour la conception d'activités collaboratives dans des MSE. Elle offre également un outil de prototypage rapide, qui peut être facilement utilisé par des non-experts, tels que des enseignants, afin de créer des activités de prise de décision collaboratives dans des MSE
Multi-surface environments (MSEs) combine several of surfaces in a variety of physical arrangements to form a seamless information space. The large surfaces, such as tabletops and wall-displays are often used as a shared space to coordinate efforts, and handheld devices, such as tablets and smartphones are regarded as personal space supporting individual tasks. MSEs have shown benefits for supporting co-located activities, especially the ones involving rich data exploration, such as complex collaborative problem-solving and decision-making activities. However, the diversity of MSE also raises questions and challenges, as different configuration and devices factors of MSE can be suited for different kinds of activities, and developing collaborative activities in MSE remains complex. This dissertation studies how MSE can support users' collaboration in general, and collaborative learning specifically. It provides insights on how the configuration and form factors of devices in MSE shape users' collaborative behaviors, and offers implications on designing collaborative activities in MSE. It also contributes with a rapid prototyping tool, which can be easily used by non-experts, such as teachers, to create collaborative decision-making activities in MSE

Published Article
The relationship between the big five personality traits, Extraversion, Agreeableness Neuroticism, Conscientiousness and Openness to Experience and deep and surface approaches to learning forms the basis of this article. The findings of a research study in this milieu will be presented to prove that earlier studies in this field have been upheld, but that an important deviation has occurred on certain levels of personality. A students way of learning implies the type of learning that is taking place. Ultimately we as lecturers want to encourage deep learning as this stimulates retention of information, important in production of students that are ready for employment.

Transition metals represent some of the first catalysts used in industrial processes and are still used today to produce many of the most needed chemicals. Adopting from ancient metallurgical techniques, it followed that the performance of these basic transition metals can be refined by adding multiple components. Since that time, improvements to these alloy catalysts has been mostly incremental due to the difficulty of producing new catalysts experimentally and a lack of fundamental understanding of the underlying physics. More recently, computational chemistry has proven itself an increasingly effective means for identifying these underlying physics. Through the use of d-band interactions of adsorbates with the surface, basic adsorption characteristics can be predicted across transition metals with limited initial information. However, although these models function well as high-level screening tools, much work is yet to be done before optimal catalysts can be comfortably designed from properties which experimentalists can directly control. This remains particularly challenging for alloy modeling, primarily due to the large number of possible atomic configurations, even for two metal systems. This work focuses on developing the methods for modeling optimal reaction properties at the surface of a transition metal alloy. Based on thermodynamic equilibrium between the surface, bulk, and gas reservoir, a model for the prediction of segregation under vacuum and adsorbate conditions can be predicted. Furthermore, by relating strain in the bulk lattice constant to the adsorption energies of varying local active sites, the optimal surface compositions can be related to bulk composition; a feature which can easily be selected for. Although useful for identifying trends across bulk composition space, these methods are limited to a small subset of active site configurations. To capture the complexity of more sophisticated processes, such as segregation, higher-timescale methods are required. Traditional computational tools are often too expensive to implement for these methods, and as such, they are usually completed with less-accurate potentials. In this work, we demonstrate that machine learning techniques have improved accuracy compared to physical potentials. We then go on to demonstrate how this improved accuracy can lead to experimentally accurate predictions of segregation.

Through the development and implementation of modified learning contexts, the current study encouraged undergraduate teacher education students to modify their approaches to learning by reducing their reliance on surface approaches and progressively adopting deeper approaches. This outcome was considered desirable because students who employed deep approaches would exit the course having achieved higher quality learning than those who relied primarily on surface approaches. It was expected that higher quality learning in a preservice teacher education program would also translate into greater self-confidence in the management of teaching tasks, leading to improvements in students� teaching self-efficacy beliefs. Altered learning contexts were developed through the application of action research methodology involving core members of the teaching team. Learning activities were designed with a focus on co-operative small-group problem-based learning, which included multiple subtasks requiring variable outcome presentation modes. Linked individual reflection was encouraged by personal learning journals and learning portfolios. Students also provided critical analyses of their own learning during the completion of tasks, from both individual and group perspectives. Assessment methods included lecturer, peer and self-assessment, depending on the nature of the learning task. Often these were integrated, so that subtasks within larger ones were assessed using combinations of methods. Learning approach theorists (Biggs, 1993a, 1999; Entwistle, 1986, 1998; Prosser & Trigwell, 1999; Ramsden, 1992, 1997) contend that learning outcomes are directly related to the learning approaches used in their development. They further contend that the approach adopted is largely a result of students� intent, which in turn, is influenced by their perception of the learning context. The present study therefore aimed to develop an integrated and pervasive course-based learning context, constructively aligned (after: Biggs, 1993a, 1996), achievable within the normal constraints of a university program, that would influence students� adoption of deep learning approaches. The cognitive processes students used in response to the altered contexts were interpreted in accordance with self-regulatory internal logic (after: Bandura, 1986, 1991b; Zimmerman, 1989, 1998b). Longitudinal quasi-experimental methods with repeated measures on non-equivalent dependent variables were applied to three cohorts of students. Cohort 1 represented the contrast group who followed a traditional program. Cohort 2 was the main treatment group to whom the modified program was presented. Cohort 3 represented a comparison group that was also presented with the modified program over a shorter period. Student data on learning approach, teaching efficacy and academic attributions were gathered from repeated administrations of the Study Process Questionnaire (Biggs, 1987b), Teacher Efficacy Scale (Gibson & Dembo, 1984) and Multidimensional-Multiattributional Causality Scale (Lefcourt, 1991). In addition, reflective journals, field observations and transcripts of interviews undertaken at the beginning and conclusion of the course, were used to clarify students� approaches to learning and their responses to program modifications. Analyses of learning approaches adopted by Cohorts 1 and 2 revealed that they both began their course predominantly using surface approaches. While students in Cohort 1 completed the course with approximately equal reliance on deep and surface approaches, students in Cohort 2 reported a predominant use of deep approaches on course completion. The relative impact of the modified learning context on students with differing approaches to learning in this cohort were further explained through qualitative data and cluster analyses. The partial replication of the study with Cohort 3, across the first three semesters of their program, produced similar effects to those obtained with Cohort 2. The analyses conducted with teaching efficacy data indicated a similar pattern of development for all cohorts. Little change in either personal or general dimensions was noted in the first half of the program, followed by strong growth in both, in the latter half. While a relationship between learning approach usage and teaching efficacy was not apparent in Cohort 1, developmental path and mediation analyses indicated that the use of deep learning approaches considerably influenced the development of personal teaching efficacy in Cohort 2. The current research suggests that value lies in the construction of learning environments, in teacher education, that enhance students� adoption of deep learning approaches. The nature of the task is complex, multifaceted and context specific, most likely requiring the development of unique solutions in each environment. Nevertheless, this research demonstrates that such solutions can be developed and applied within the prevailing constraints of pre-existing course structures.

Through the development and implementation of modified learning contexts, the current study encouraged undergraduate teacher education students to modify their approaches to learning by reducing their reliance on surface approaches and progressively adopting deeper approaches. This outcome was considered desirable because students who employed deep approaches would exit the course having achieved higher quality learning than those who relied primarily on surface approaches. It was expected that higher quality learning in a preservice teacher education program would also translate into greater self-confidence in the management of teaching tasks, leading to improvements in students� teaching self-efficacy beliefs. Altered learning contexts were developed through the application of action research methodology involving core members of the teaching team. Learning activities were designed with a focus on co-operative small-group problem-based learning, which included multiple subtasks requiring variable outcome presentation modes. Linked individual reflection was encouraged by personal learning journals and learning portfolios. Students also provided critical analyses of their own learning during the completion of tasks, from both individual and group perspectives. Assessment methods included lecturer, peer and self-assessment, depending on the nature of the learning task. Often these were integrated, so that subtasks within larger ones were assessed using combinations of methods. Learning approach theorists (Biggs, 1993a, 1999; Entwistle, 1986, 1998; Prosser & Trigwell, 1999; Ramsden, 1992, 1997) contend that learning outcomes are directly related to the learning approaches used in their development. They further contend that the approach adopted is largely a result of students� intent, which in turn, is influenced by their perception of the learning context. The present study therefore aimed to develop an integrated and pervasive course-based learning context, constructively aligned (after: Biggs, 1993a, 1996), achievable within the normal constraints of a university program, that would influence students� adoption of deep learning approaches. The cognitive processes students used in response to the altered contexts were interpreted in accordance with self-regulatory internal logic (after: Bandura, 1986, 1991b; Zimmerman, 1989, 1998b). Longitudinal quasi-experimental methods with repeated measures on non-equivalent dependent variables were applied to three cohorts of students. Cohort 1 represented the contrast group who followed a traditional program. Cohort 2 was the main treatment group to whom the modified program was presented. Cohort 3 represented a comparison group that was also presented with the modified program over a shorter period. Student data on learning approach, teaching efficacy and academic attributions were gathered from repeated administrations of the Study Process Questionnaire (Biggs, 1987b), Teacher Efficacy Scale (Gibson & Dembo, 1984) and Multidimensional-Multiattributional Causality Scale (Lefcourt, 1991). In addition, reflective journals, field observations and transcripts of interviews undertaken at the beginning and conclusion of the course, were used to clarify students� approaches to learning and their responses to program modifications. Analyses of learning approaches adopted by Cohorts 1 and 2 revealed that they both began their course predominantly using surface approaches. While students in Cohort 1 completed the course with approximately equal reliance on deep and surface approaches, students in Cohort 2 reported a predominant use of deep approaches on course completion. The relative impact of the modified learning context on students with differing approaches to learning in this cohort were further explained through qualitative data and cluster analyses. The partial replication of the study with Cohort 3, across the first three semesters of their program, produced similar effects to those obtained with Cohort 2. The analyses conducted with teaching efficacy data indicated a similar pattern of development for all cohorts. Little change in either personal or general dimensions was noted in the first half of the program, followed by strong growth in both, in the latter half. While a relationship between learning approach usage and teaching efficacy was not apparent in Cohort 1, developmental path and mediation analyses indicated that the use of deep learning approaches considerably influenced the development of personal teaching efficacy in Cohort 2. The current research suggests that value lies in the construction of learning environments, in teacher education, that enhance students� adoption of deep learning approaches. The nature of the task is complex, multifaceted and context specific, most likely requiring the development of unique solutions in each environment. Nevertheless, this research demonstrates that such solutions can be developed and applied within the prevailing constraints of pre-existing course structures.

Le domaine des véhicules autonomes a suscité un grand intérêt ces dernières années. Afin de garantir au passager une expérience sûre et confortable sur les véhicules autonomes, des systèmes d'obstacles avancés doivent être mis en œuvre. Bien que les solutions actuelles de détection d'obstacles aient montré de bonnes performances, elles doivent être encore améliorées pour une sécurité accrue des véhicules autonomes sur route, de jour comme de nuit. En particulier, les véhicules autonomes dans la vie réelle peuvent rencontrer de la glace, de la neige ou des flaques d'eau, qui peuvent être la cause de collisions graves et d'accidents de la circulation. Les systèmes de détection doivent donc permettre de détecter les changements d'état de la route pour anticiper la réaction du véhicule et/ou désactiver les fonctions automatisées. L'objectif de cette thèse est de proposer un système pour les véhicules autonomes afin de détecter les conditions de chaussée induites par la météo. Après une étude approfondie de l'état de l'art, un système proche infrarouge (NIR) basé sur des LED et un système d'apprentissage automatique sont proposés pour la détection diurne et nocturne. Le système NIR a été conçu puis validé expérimentalement et, les spécifications techniques du système ont été définies. Le système d'apprentissage automatique est de plus proposé comme solution complémentaire au système NIR. Différents modèles d'apprentissage ont été testés et comparés en termes de performance. Enfin, les résultats sont discutés et une combinaison des deux systèmes est proposée afin de garantir une performance accrue pour la reconnaissance des conditions de route
The field of autonomous vehicles has aroused great interest in recent years. In order to ensure the passenger to get a safe and comfortable experience on autonomous vehicles, advanced obstacle systems have to be implemented. Although current solutions for detecting obstacles have shown quite good performances, they have to be improved for an increased safety of autonomous vehicles on road, both in day-time and night-time conditions. In particular, autonomous vehicles in real life may encounter ice, snow or water puddles, which may be the cause of severe crashes and traffic accidents. The detection systems must hence allow detecting changes in road conditions to anticipate the vehicle reaction and/or deactivate the automated functions. The aim of this thesis is to propose a system implemented on the autonomous vehicles in order to detect the road surface conditions induced by the weather. After deep investigation of the state of art, a near infrared (NIR) system based on LEDs and a machine learning system were proposed for daytime and night-time detection. The NIR systems with three LEDs were investigated with experimental validations. In addition, the specifications of the NIR systems are carefully discussed. Furthermore, the machine learning system is proposed as a supplementary system. The performance of different models is compared in terms of classification accuracy and model complexity. Finally, the results are discussed and a combination of the two systems is proposed

The relationship between sEMG signals and muscle force, and associated joint torque, is an object of study for clinical applications such as rehabilitation robotics and commercial applications as wearable motion control devices. The information type and quality obtained by sEMG can impact the classification and prediction accuracy of ankle joint torque. In this thesis project, HD-sEMG based data was collected together with ankle joint torque measurements from 5 subjects during MVIC of plantarflexors and dorsiflexors. Machine learning approaches ideally suited for nonlinear regression tasks, such as MLP and LSTM, have been implemented and evaluated to best predict joint torque profiles given extracted features from sEMG data. An evaluation of machine learning performances using HD-sEMG data over bipolar sEMG data has been conducted in intra-session, inter-subjective and intra-subjective study cases.

The increasing demand for higher efficiency and lower environmental impact of transmissions, used in automotive and wind energy industries has created a need for more advanced technical solutions to fulfil those requirements. Condition monitoring plays an important role in the transmission life cycle, saving resources and time. Recently condition monitoring, using machine learning has shifted from reactive to proactive action, predicting minor faults before they become significant. This thesis intends to develop a methodology that can be used to predict faults like pitting initiation, before propagating in FZG test rig, available at KTH Machine Design department. Standard sensor measurements already available like temperature, rotation speed and torque are used in this project. Four kinds of gears were used, two made of wrought, and two – of powder metal steel, each with ground or superfinish surface. After a literature review about pitting fatigue, condition indicators for these failures and machine learning were done, a statistical analysis was done, to see how the transmission behaves during testing and to have comparison material, helpful when having machine learning results. Two machine learning models, Decision Tree and Support Vector Machine were selected and trained in two combinations, either with Root Mean Square only, or with Crest Factor, Standard Deviation and Kurtosis in addition. As a result, 64 models were trained, 32 for all tests and another 32 to investigate two particular tests due to a longer pitting propagation period. New condition indicators like Standard Deviation and Signal – to – noise ratio was calculated to get more nuanced trends than just using one measurement to monitor the gearbox behavior. After comparing with the results from statistical analysis and previously done tooth profile measurements, it was concluded that the new indicators could indicate the change in gearbox operation before the first pitting initiation is detected, using tooth profile measurement.

Modelling surface runoff can be beneficial to operations within many fields, such as agriculture planning, flood and drought risk assessment, and water resource management. In this study, we built a data-driven model that can reproduce monthly surface runoff at a 4-km grid network covering 13 watersheds in the Chesapeake Bay area. We used a random forest algorithm to build the model, where monthly precipitation, temperature, land cover, and topographic data were used as predictors, and monthly surface runoff generated by the SWAT hydrological model was used as the response. A sub-model was developed for each of 12 monthly surface runoff estimates, independent of one another. Accuracy statistics and variable importance measures from the random forest algorithm reveal that precipitation was the most important variable to the model, but including climatological data from multiple months as predictors significantly improves the model performance. Using 3-month climatological, land cover, and DEM derivatives from 40% of the 4-km grids as the training dataset, our model successfully predicted surface runoff for the remaining 60% of the grids (mean R2 (RMSE) for the 12 monthly models is 0.83 (6.60 mm)). The lowest R2 was associated with the model for August, when the surface runoff values are least in a year. In all studied watersheds, the highest predictive errors were found within the watershed with greatest topographic complexity, for which the model tended to underestimate surface runoff. For the other 12 watersheds studied, the data-driven model produced smaller and more spatially consistent predictive errors.
Master of Science

This thesis investigates a proposition that students learning from computer assisted learning (CAL) programs will be constrained to surface processing of information unless generative learning strategies are incorporated into the CAL program design. The specific generative learning recommended in the proposition include summarising, outlining, analysis of key ideas, and cognitive mapping. Studies of student learning in tertiary education reveal that surface learning is not conducive to academic success, and that deep learning is more effective. As very few CAL programs use generative learning strategies, the suggestion that students will be prevented from deep learning by CAL design is important if CAL programs are to be used in tertiary education. Two cohorts of students for whom the use of a CAL program was a part of their course unit requirements were studied to evaluate the depth of their consequent learning. No experimental controls were applied, so that the students' use of the CAL program could be observed within their normal learning environment. Neither program incorporated generative learning strategies into the program design. All research instruments used were external to the CAL programs to provide an independent perspective on the learning outcome. The Structure of Observed Leaming Outcomes (SOLO) taxonomy was applied to open-ended questions on the topics of the CAL programs to assess depth of learning. Performance as assessesd by SOLO scores was compared with the students' habitual approach to learning as indicated by the Study Process Questionnaire (SPQ) to assess the effect of the CAL program on the students' normal approaches to learning. Comparison between SPQ and SOLO scores indicated that many students did not perform in accordance with their habitual approach as indicated by the SPQ. To clarify the influence of the CAL program on performance, questions were developed to evaluate the students' use of the programs. These were administered by interview or by a questionnaire. These data were subjected to a thematic content analysis. The findings revealed that many of the students who had confidence in the CAL program performed at a deep level despite the lack of generative learning strategies in the program, and that the implementation of the programs is a major influence on student performance. Students who saw the program as an information source rather than as an interactive learning opportunity did not perform at a deep level, and students who could not see the reason for using the program performed poorly. Recommendations for CAL design include the need for designers to provide guidance for students to encourage a strategic approach to using the CAL program, and a cognitive approach to the implementation of CAL programs.

Reconstruire des champs géophysiques à partir d'observations bruitées et partielles est un problème classique bien étudié dans la littérature. L'assimilation de données est une méthode populaire pour aborder ce problème, et se fait par l'utilisation de techniques classiques, comme le filtrage de Kalman d’ensemble ou des filtres particulaires qui procèdent à une évaluation online du modèle physique afin de fournir une prévision de l'état. La performance de l'assimilation de données dépend alors fortement de du modèle physique. En revanche, la quantité de données d'observation et de simulation a augmenté rapidement au cours des dernières années. Cette thèse traite l'assimilation de données d'une manière data-driven et ce, sans avoir accès aux équations explicites du modèle. Nous avons développé et évalué l'assimilation des données par analogues (AnDA), qui combine la méthode des analogues et des méthodes de filtrage stochastiques (filtres Kalman, filtres à particules, chaînes de Markov cachées). Des applications aux modèles chaotiques simplifiés et à des études de cas de télédétection réelle (température de surface de lamer, anomalies du niveau de la mer), nous démontrons la pertinence d'AnDA pour l'interpolation de données manquantes des systèmes dynamiques non linéaires et à haute dimension à partir d'observations irrégulières et bruyantes.Motivé par l'essor du machine learning récemment, la dernière partie de cette thèse est consacrée à l'élaboration de modèles deep learning pour la détection et de tourbillons océaniques à partir de données de sources multiples et/ou multi temporelles (ex: SST-SSH), l'objectif général étant de surpasser les approches dites expertes
Reconstructing geophysical fields from noisy and partial remote sensing observations is a classical problem well studied in the literature. Data assimilation is one class of popular methods to address this issue, and is done through the use of classical stochastic filtering techniques, such as ensemble Kalman or particle filters and smoothers. They proceed by an online evaluation of the physical modelin order to provide a forecast for the state. Therefore, the performanceof data assimilation heavily relies on the definition of the physical model. In contrast, the amount of observation and simulation data has grown very quickly in the last decades. This thesis focuses on performing data assimilation in a data-driven way and this without having access to explicit model equations. The main contribution of this thesis lies in developing and evaluating the Analog Data Assimilation(AnDA), which combines analog methods (nearest neighbors search) and stochastic filtering methods (Kalman filters, particle filters, Hidden Markov Models). Through applications to both simplified chaotic models and real ocean remote sensing case-studies (sea surface temperature, along-track sea level anomalies), we demonstrate the relevance of AnDA for missing data interpolation of nonlinear and high dimensional dynamical systems from irregularly-sampled and noisy observations. Driven by the rise of machine learning in the recent years, the last part of this thesis is dedicated to the development of deep learning models for the detection and tracking of ocean eddies from multi-source and/or multi-temporal data (e.g., SST-SSH), the general objective being to outperform expert-based approaches

La Génération Automatique de Langue Naturelle vise à produire des textes dans une langue humaine à partir d'un ensemble de données non-linguistiques. Elle comprend généralement trois sous-tâches principales: (i) sélection et organisation d'un sous-ensemble des données d'entrée; ii) détermination des mots à utiliser pour verbaliser les données d'entrée; et (iii) regroupement de ces mots en un texte en langue naturelle. La dernière sous-tâche est connue comme la tâche de Réalisation de Surface (RS). Dans ma thèse, j'étudie la tâche de RS quand les données d'entrée sont extraites de Bases de Connaissances (BC). Je présente deux nouvelles approches pour la réalisation de surface à partir de bases de connaissances: une approche supervisée et une approche faiblement supervisée. Dans l'approche supervisée, je présente une méthode basée sur des corpus pour induire une grammaire à partir d'un corpus parallèle de textes et de données. Je montre que la grammaire induite est compacte et suffisamment générale pour traiter les données de test. Dans l'approche faiblement supervisée, j'explore une méthode pour la réalisation de surface à partir de données extraites d'une BC qui ne requière pas de corpus parallèle. À la place, je construis un corpus de textes liés au domaine et l'utilise pour identifier les lexicalisations possibles des symboles de la BC et leurs modes de verbalisation. J'évalue les phrases générées et analyse les questions relatives à l'apprentissage à partir de corpus non-alignés. Dans chacune de ces approches, les méthodes proposées sont génériques et peuvent être facilement adaptées pour une entrée à partir d'autres ontologies
Natural Language Generation is the task of automatically producing natural language text to describe information present in non-linguistic data. It involves three main subtasks: (i) selecting the relevant portion of input data; (ii) determining the words that will be used to verbalise the selected data; and (iii) mapping these words into natural language text. The latter task is known as Surface Realisation (SR). In my thesis, I study the SR task in the context of input data coming from Knowledge Bases (KB). I present two novel approaches to surface realisation from knowledge bases: a supervised approach and a weakly supervised approach. In the first, supervised, approach, I present a corpus-based method for inducing a Feature Based Lexicalized Tree Adjoining Grammar from a parallel corpus of text and data. I show that the induced grammar is compact and generalises well over the test data yielding results that are close to those produced by a handcrafted symbolic approach and which outperform an alternative statistical approach. In the weakly supervised approach, I explore a method for surface realisation from KB data which does not require a parallel corpus. Instead, I build a corpus from heterogeneous sources of domain-related text and use it to identify possible lexicalisations of KB symbols and their verbalisation patterns. I evaluate the output sentences and analyse the issues relevant to learning from non-parallel corpora. In both these approaches, the proposed methods are generic and can be easily adapted for input from other ontologies for which a parallel/non-parallel corpora exists

There is an abundance of literature on research about teaching and learning in the tertiary education sector. Within this body of literature there lies the field of andragogy that focuses on the facilitation of adult learning. Although adults and youth (those aged 17-24 years) often share common learning environments, where several principles of andragogy are practiced, there is no evidence of research to ascertain whether such principles apply to youth learners. The primary purpose of this thesis was to examine whether youth learned like adults. Three characteristics that most adult learners share are: a deep approach to learning, an andragogical orientation to study and a high level of readiness for self-directed learning. The thesis firstly investigated whether youth learners also shared these characteristics and then explored the factors that contributed to their learning. Altogether, 450 youth who were enrolled in courses offered by universities and Technical and Further Education (TAFE) fustitutes completed three survey questionnaires. The Study Process Questionnaire gathered data about their learning approaches, the Student Orientation Questionnaire collected data about their study orientation and the Learning Preference Assessment questionnaire informed about their level of readiness for self-directed learning. The quantitative data from the survey were analysed using the SPSS computer software. Two analytical models were developed to ascertain whether youth learned like adults. The findings from both analytical models concluded that most youth did not learn like adult learners. Unlike adults, most youth learners had a surface approach to learning, a preference for pedagogical as well as andragogical orientations to study and low levels of readiness for self-directed learning. The survey results showed that youth preferred only the 'feel good' aspects of andragogy. Focus groups were arranged with volunteer youths who had participated in the survey. A proforma was used to explore youths' perspectives of the factors that contributed to their learning as illustrated by the survey results. Youths' teachers were interviewed to gather their perspectives of factors that contributed to youths' learning. A number of factors were identified during the focus groups and interviews. An analytical framework was developed to examine the factors. Two major themes emerged from the data: lifeworld and formal learning environment. Factors within each of these appeared to influence youths' decision making about how they chose to learn. Factors within youths' lifeworld related to their role conflict, expectancy valence and personological attributes. Institutional systems, teachers and their practices, decision making in system-related matters, and opportunities for self-directed learning, critical thinking and reflective thinking were perceived to be the main factors within the formal learning environment that contributed to youth's learning. The findings from the survey and focus group data were used to draw a profile of youth learners in terms of their priorities, motivation and learning attributes. The findings were also used for discussions relating to the six principles of adult learning. A set of principles for practice and skilling for higher learning were suggested for youth learners. Based on the results of this exploratory study, the following conclusion about youth learning was proposed: Most youth use a surface approach to learning, are at Stage 2 of their learning on an orthogonal scale and have low level of readiness for self-directed learning. Most youth seem to appreciate a relational level of understanding rather than abstract thinking. Youths' learning could be better facilitated using Kolb's learning theory such that their learning could begin with concrete experience followed by reflective observation and then abstract conceptualisation. A directive, but highly supportive approach where the facilitator plays the role of a motivator and guide, is recommended for youth learners.

Tsunami and storm surge are two of the main destructive and costly natural hazards faced by coastal communities around the world. To enhance coastal resilience and to develop effective risk management strategies, accurate and efficient tsunami and storm surge prediction models are needed. However, existing physics-based numerical models have the disadvantage of being difficult to satisfy both accuracy and efficiency at the same time. In this dissertation, several surrogate models are developed using statistical and machine learning techniques that can rapidly predict a tsunami and storm surge without substantial loss of accuracy, with respect to high-fidelity physics-based models. First, a tsunami run-up response function (TRRF) model is developed that can rapidly predict a tsunami run-up distribution from earthquake fault parameters. This new surrogate modeling approach reduces the number of simulations required to build a surrogate model by separately modeling the leading order contribution and the residual part of the tsunami run-up distribution. Secondly, a TRRF-based inversion (TRRF-INV) model is developed that can infer a tsunami source and its impact from tsunami run-up records. Since this new tsunami inversion model is based on the TRRF model, it can perform a large number of tsunami forward simulations in tsunami inversion modeling, which is impossible with physics-based models. And lastly, a one-dimensional convolutional neural network combined with principal component analysis and k-means clustering (C1PKNet) model is developed that can rapidly predict the peak storm surge from tropical cyclone track time series. Because the C1PKNet model uses the tropical cyclone track time series, it has the advantage of being able to predict more diverse tropical cyclone scenarios than the existing surrogate models that rely on a tropical cyclone condition at one moment (usually at or near landfall). The surrogate models developed in this dissertation have the potential to save lives, mitigate coastal hazard damage, and promote resilient coastal communities.
Doctor of Philosophy
Tsunami and storm surge can cause extensive damage to coastal communities; to reduce this damage, accurate and fast computer models are needed that can predict the water level change caused by these coastal hazards. The problem is that existing physics-based computer models are either accurate but slow or less accurate but fast. In this dissertation, three new computer models are developed using statistical and machine learning techniques that can rapidly predict a tsunami and storm surge without substantial loss of accuracy compared to the accurate physics-based computer models. Three computer models are as follows: (1) A computer model that can rapidly predict the maximum ground elevation wetted by the tsunami along the coastline from earthquake information, (2) A computer model that can reversely predict a tsunami source and its impact from the observations of the maximum ground elevation wetted by the tsunami, (3) A computer model that can rapidly predict peak storm surges across a wide range of coastal areas from the tropical cyclone's track position over time. These new computer models have the potential to improve forecasting capabilities, advance understanding of historical tsunami and storm surge events, and lead to better preparedness plans for possible future tsunamis and storm surges.

This study aims to predict the binding affinity of the PCP-like compounds by means of molecular hydrophobicity. Molecular hydrophobicity is an important property which affects the binding affinity of molecules. The values of molecular hydrophobicity of molecules are obtained on three-dimensional coordinate system. Our aim is to reduce the number of points on the hydrophobicity surface of the molecules. This is modeled by using self organizing maps (SOM) and k-means clustering. The feature sets obtained from SOM and k-means clustering are used in order to predict binding affinity of molecules individually. Support vector regression and partial least squares regression are used for prediction.

Content: Identification and classification of leather species becomes valuable and necessary due to concerns regarding consumer protection, product counterfeiting, and dispute settlement in the leather industry. Identification and classification of leather into species is carried out by histological examination or molecular analysis based on DNA. Manual method requires expertise, training and experience, and due to involvement of human judgment disputes are inevitable thus a need to automate the leather species identification. In the present investigation, an attempt has been made to automate leather species identification using machine learning techniques. A novel non-destructive leather species identification algorithm is proposed for the identification of cow, buffalo, goat and sheep leathers. Hair pore pattern was segmented efficiently using k-means clustering algorithm Significant features representing the unique characteristics of each species such as no.of hair pores, pore density, percent porosity, shape of the pores etc., were extracted. The generated features were used for training the Random forest classifier. Experimental results on the leather species image library database achieved an accuracy of 87 % using random forest as classifier, confirming the potentials of using the proposed system for automatic leather species classification. Take-Away: Novel technique to identify leather species Non destructive method Machine learning algorithms to automate leather species identification

Cette thèse sur l’analyse multimodale du comportementale des consommateurs dans une surface de vente se décline en trois problématiques : 1) la reconnaissance des caractéristiques bas-niveau que sont la tranche d’âge et le genre, 2) l’analyse des caractéristiques moyen-niveau telles que le degré ou la classe d’indécision 3) l’identification des caractéristiques hauts niveaux que sont les types d’achat. Les données proviennent de caméra, de capteurs oculométriques et de capteurs de position. L’identification du genre et de la tranche d’âge s’effectue sur des images de visages. Notre première contribution est de proposer un modèle d'identification du genre et de la tranche d'âge, en se basant sur des descripteurs de texture sur la partie centrale du visage. Nous mettons en exergue une corrélation entre la tranche d'âge du sujet et le degré de difficulté à identifier son genre. Cette corrélation légitime la segmentation marketing de la population en tranches d'âge. Notre seconde contribution concerne la seconde problématique. En effet, nous proposons une analyse prédictive, et non plus descriptive, du degré d'indécision. Nous utilisons pour ce faire des descripteurs oculométriques et de préhension, ainsi que les machines à vecteurs de support. Notre troisième contribution concerne l'analyse du type d'achat sur des données oculométriques. Tout comme pour l'analyse du degré d'indécision, nous proposons un modèle prédictif. Nous mettons en exergue le facteur temps, important dans tout acte d'achat.Cette thèse a été initiée au sein du projet ANR ORIGAMI2 : Observation du Regard et Interprétation du Geste pour une Analyse Marketing non Intrusive
This thesis about multimodal analysis of customer behavior in a selling area falls into three issues: 1) the identification of low level characteristics such as age band and gender, 2) the analysis of middle level characteristics such as the indecisiveness degree or the indecisiveness class, 3) the identification of purchasing acts. Data come from cameras, eye-tracking sensors and infrared position sensors. The identification of gender and age band is made with images of faces. Our first contribution is proposing a model for the identification of the gender and the age band, based on texture descriptors on the middle third of the face. We point out a correlation between the age of the subject and the difficulty to identify his gender. This correlation legitimizes the segmentation of the population by marketing managers into age bands. Our second contribution deals with the second issue. Indeed, we propose a predictive analysis of the indecisiveness degree of the customer, instead of descriptive analysis. We use eye-tracking descriptors, gesture descriptors and support vector machines. Our third contribution deals with the analysis of purchasing acts based on eye tracking data. As for the analysis of the indecisiveness degree, we propose a predictive model. We emphasize the time factor, which is an important factor in the purchasing act.This thesis was initiated within the behavioral marketing project ORIGAMI2: observation of gaze and interpretation of gesture for a non-intrusive marketing analysis (Observation du Regard et Interprétation du Geste pour Analyse Marketing non-Intrusive)

Los juegos serios colaborativos tienen un impacto positivo en el comportamiento y el aprendizaje, pero siguen desarrollándose para plataformas tecnológicas tradicionales como videoconsolas y ordenadores de sobremesa o portátiles, los cuales han sido identificados como sub-óptimos para niños en diversos estudios. En su lugar, el uso de dispositivos móviles como tabletas y teléfonos inteligentes presenta diversas ventajas: son económicamente asequibles, están ampliamente distribuidos, y pueden ser transportados, lo cual permite la actividad física y poder iniciar un juego sin necesitar que los usuarios se trasladen a una localización fija, especialmente dedicada para tal fin. Además, combinar varios de estos dispositivos y coordinar la interacción entre ellos en lo que se denomina Entorno Multi-Pantalla (EMP) proporciona beneficios adicionales para la colaboración tales como una mayor escalabilidad, conciencia del espacio de trabajo, paralelismo y fluidez de las interacciones. La interacción en estos entornos multi-tableta es por tanto un aspecto crítico. Los dispositivos móviles están diseñados para ser interactuados mediante el toque de los dedos principalmente, lo cual es muy sencillo y directo, pero está normalmente limitado a la pequeña dimensión de las pantallas, lo que puede conllevar la oclusión de la pantalla y la infrautilización del espacio periférico. Por esta razón, esta tesis se centra en la exploración de otro mecanismo de interacción que puede complementar al táctil: interacciones tangibles alrededor del dispositivo. Las interacciones tangibles están basadas en la manipulación de objetos físicos, lo que presenta un valor adicional en la educación de los niños puesto que resuena con los manipulativos educativos tradicionales y permite la exploración del mundo físico. Por otra parte, la explotación del espacio que envuelve a las pantallas tiene diversos beneficios adicionales para actividades educativas colaborativas: reducida oclusión de la pantalla (lo cual puede incrementar la conciencia del espacio de trabajo), el uso de objetos tangibles como contenedores de información digital que puede ser transportada de forma continua entre dispositivos, y la identificación de un determinado estudiante a través de la codificación de su ID en un operador tangible (lo cual facilita el seguimiento de sus acciones y progreso durante el juego). Esta tesis describe dos enfoques distintos para construir juegos educativos colaborativos en EMPs utilizando interacciones tangibles alrededor de los dispositivos. Una, denominada MarkAirs, es una solución óptica aérea que no necesita ningún hardware adicional aparte de las tabletas excepto diversas tarjetas de cartón impresas. La otra, Tangibot, introduce un robot tangiblemente controlado y otro atrezo físico en el entorno, y se basa en tecnología RFID. Ambas interacciones son respectivamente evaluadas, y se observa que MarkAirs es usable y poco exigente tanto para adultos como para niños, y que se pueden realizar con éxito gestos de grano fino encima de las tabletas con ella. Además, al aplicarse en juegos colaborativos, puede ayudar a reducir la oclusión de las pantallas y la interferencia entre las distintas acciones de los usuarios, lo cual es un problema que puede surgir en este tipo de escenarios cuando solamente se dispone de interacciones táctiles. Se evalúa un juego educativo colaborativo con MarkAirs con niños de educación primaria, y se concluye que este mecanismo es capaz de crear experiencias de aprendizaje colaborativo y de presentar un valor añadido en términos de experiencia de usuario, aunque no en eficiencia. Con respecto a Tangibot, se muestra que controlar colaborativamente un robot móvil mediante unas palas tangibles con cierta precisión es factible para niños a partir de los tres años de edad, e incluso para personas mayores con un deterioro cognitivo leve. Además, proporciona una experiencia divertida
Collaborative serious games have a positive impact on behavior and learning, but the majority are still being developed for traditional technological platforms, e.g., video consoles and desktop/laptop computers, which have been deemed suboptimal for children by several studies. Instead, the use of handheld devices such as tablets and smartphones presents several advantages: they are affordable, very widespread, and mobile---which enables physical activity and being able to engage in a game without requiring users to gather around a fixed, dedicated, location. Plus, combining several of these devices and coordinating interactions across them in what is called a Multi-Display Environment (MDE) brings on additional benefits to collaboration like higher scalability, awareness, parallelism, and fluidity of the interaction. How to interact with these multi-tablet environments is therefore a critical issue. Mobile devices are designed to be interacted mainly via touch, which is very straightforward but usually limited to the small area of the displays, which can lead to the occlusion of the screen and the underuse of the peripheral space. For this reason, this thesis focuses on the exploration of another interaction mechanism that can complement touch: tangible around-device interactions. Tangible interactions are based on the manipulation of physical objects, which have an added value in childhood education as they resonate with traditional learning manipulatives and enable the exploration of the physical world. On the other hand, the exploitation of the space surrounding the displays has several potential benefits for collaborative-learning activities: reduced on-screen occlusion (which may increase workspace awareness), the use of tangible objects as containers of digital information that can be seamlessly moved across devices, and the identification of a given student through the encoding of their ID in a tangible manipulator (which facilitates the tracking of their actions and progress throughout the game). This thesis describes two different approaches to build collaborative-learning games for MDEs using tangible around-device interactions. One, called MarkAirs, is a mid-air optical solution relying on no additional hardware besides the tablets except for several cardboard printed cards. The other, Tangibot, introduces a tangible-mediated robot and other physical props in the environment and is based on RFID technology. Both interactions are respectively evaluated, and it is observed that MarkAirs is usable and undemanding both for adults and for children, and that fine-grained gestures above the tablets can be successfully conducted with it. Also, when applied to collaborative games, it can help reduce screen occlusion and interference among the different users' actions, which is a problem that may arise in such settings when only touch interactions are available. A collaborative learning game with MarkAirs is evaluated with primary school children, revealing this mechanism as capable of creating collaborative learning experiences and presenting an added value in user experience, although not in performance. With respect to Tangibot, we show how collaboratively controlling a mobile robot with tangible paddles and achieving certain precision with it is feasible for children from 3 years of age, and even for elderly people with mild cognitive impairment. Furthermore, it provides a fun experience for children and maintains them in a constant state of flow.
Els jocs seriosos col·laboratius tenen un impacte positiu en el comportament i l'aprenentatge, però continuen sent desenvolupats per a plataformes tecnològiques tradicionals com videoconsoles i ordinadors de sobretaula o portàtils, els quals han sigut identificats com sub-òptims per a xiquets en diversos estudis. D'altra banda, l'ús de dispositius mòbils com ara tabletes i telèfons intel·ligents presenta diversos avantatges: són econòmicament assequibles, estan àmpliament distribuïts i poden ser transportats, la qual cosa permet l'activitat física i poder iniciar un joc sense necessitat de què els usuaris es traslladen a una localització fixa i especialment dedicada per a eixa finalitat. A més, combinar diversos d'estos dispositius i coordinar la interacció entre ells en el que es denomina Entorn Multi-Pantalla (EMP) proporciona beneficis addicionals per a la col·laboració tals com una major escalabilitat, consciència de l'espai de treball, paral·lelisme i fluïdesa de les interaccions. La interacció amb estos entorns multi-tableta és per tant crítica. Els dispositius mòbils estan dissenyats per a ser interactuats mitjançant tocs de dit principalment, mecanisme molt senzill i directe, però està normalment limitat a la reduïda dimensió de les pantalles, cosa que pot ocasionar l'oclusió de la pantalla i la infrautilització de l'espai perifèric. Per aquesta raó, la present tesi se centra en l'exploració d'un altre mecanisme d'interacció que pot complementar al tàctil: interaccions tangible al voltant dels dispositius. Les interaccions tangibles estan basades en la manipulació d'objectes físics, cosa que presenta un valor addicional en l'educació dels xiquets ja que ressona amb els manipulatius tradicionals i permet l'exploració del món físic. D'altra banda, l'explotació de l'espai que envolta a les pantalles té diversos beneficis addicionals per a activitats educatives col·laboratives: reduïda oclusió de la pantalla (la qual cosa pot incrementar la consciència de l'espai de treball), l'ús d'objectes tangibles com a contenidors d'informació digital que pot ser transportada de forma continua entre dispositius, i la identificació d'un estudiant determinat a través de la codificació de la seua identitat en un operador tangible (cosa que facilita el seguiment de les seues accions i progrés durant el joc). Aquesta tesi descriu dos enfocaments distints per a construir jocs educatius col·laboratius en EMPs utilitzant interaccions tangibles al voltant dels dispositius. Una, denominada MarkAirs, és una solució òptica aèria que no precisa de cap maquinari addicional a banda de les tabletes, exceptuant diverses targetes de cartró impreses. L'altra, Tangibot, introdueix un robot controlat tangiblement i attrezzo físic addicional en l'entorn, i es basa en tecnologia RFID. Ambdues interaccions són avaluades respectivament, i s'observa que MarkAirs és usable i poc exigent tant per a adults com per a xiquets, i que es poden realitzar gestos de granularitat fina dalt de les tabletes amb ella. A més a més, en aplicar-se a jocs col·laboratius, pot ajudar a reduir l'oclusió de les pantalles i la interferència entre les distintes accions dels usuaris, problema que pot aparèixer en este tipus d'escenaris quan solament es disposa d'interaccions tàctils. S'avalua un joc educatiu col·laboratiu amb MarkAirs amb xiquets d'educació primària, i es conclou que aquest mecanisme és capaç de crear experiències d'aprenentatge col·laboratiu i de presentar un valor afegit en termes d'experiència d'usuari, tot i que no en eficiència. Respecte a Tangibot, es mostra que controlar conjuntament un robot mòbil mitjançant unes pales tangibles amb certa precisió és factible per a xiquets a partir de tres anys i inclús per a persones majors amb un lleu deteriorament cognitiu. A més, proporciona una experiència divertida per als xiquets i els manté en un estat constant de flow.
García Sanjuan, F. (2018). CREAME: CReation of Educative Affordable Multi-surface Environments [Tesis doctoral no publicada]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/101942
TESIS

This thesis presents the results of a study that embraces and tests Entwistle's theory of deep and surface approaches in relation to students’ interaction with a digital learning environment for mathematics, in real conditions, during tutorial sessions. In contrast to most of the work in the field that seeks ways of adapting a system to students’ specific learning styles, the aim is to find ways to support tutors and researchers to identify students’ prominent approach in order to ultimately encourage the adoption of a deep approach to studying while discouraging a surface approach. To achieve this aim there is an in-depth examination of the relationship between the various scales and subscales of the Approaches and Study Skills Inventory for Students (ASSIST) and metrics occurring from the interaction in the digital learning environment ActiveMath. Furthermore, the potential influence of students’ prior knowledge in mathematics in “deep” and “surface” models is discussed. The results point to insights for tutors regarding identifying students’ deep and surface approaches from their interaction with the digital learning environment; suggestions regarding the design of features that encourage a deep approach to studying; and methodological recommendations for researchers regarding future studies which can help to distinguish further deep and surface approaches and to examine them in similar or different educational settings.

Deformation twinning in Magnesium alloys both facilitates slip and forms sites for failure. Currently, basic studies of twinning in Mg are facilitated by electron backscatter diffraction (EBSD) which is able to extract a myriad of information relating to crystalline microstructures. Although much information is available via EBSD, various problems relating to deformation twinning have not been solved. This dissertation provides new insights into deformation twinning in Mg alloys, with particular focus on AZ31. These insights were gained through the development of new EBSD and related machine learning tools that extract more information beyond what is currently accessed.The first tool relating to characterization of deformed and twinned materials focuses on surface topography crack detection. The intensity map across EBSD images contains vital information that can be used to detect evolution of surface roughness and crack formation, which typically occurs at twin boundaries. The method of topography recovery resulted in reconstruction errors as low as 2% over a 500 μm length. The method was then applied to a 3 μm x 3 μm area of twinned Tantalum which experienced topographic alterations. The topography of Ta correlated with other measured changes in the microstructure. Additionally, EBSD images were used to identify the presence of cracks in Nickel microstructures. Several cracks were identified on the Ni specimen, demonstrating that cracks as thin as 34 nm could be measured.A further EBSD based tool developed for this study was used to identify thin compression twins in Mg; these are often missed in a traditional EBSD scan due to their size relative to the electron probe. This tool takes advantage of crystallographic relationships that exist between parent and twinned grains; common planes that exist in both grains lead to bands of consistent intensity as a scan crosses a twin. Hence, twin boundaries in a microstructure can be recognized, even when they are associated with thin twins. Proof of concept was performed on known twins in Inconel 600, Tantalum, and Magnesium AZ31. This method was then used to search for undetected twins in a Mg AZ31 structure, revealing nearly double the number of twins compared with those initially measured by standard procedures.To uncover the driving forces behind deformation twinning in Mg, a machine learning framework was developed to leverage all of the data available from EBSD and use that to create a physics based models of twin nucleation and growth. The resultant models for nucleation and growth were measured to be up to 86.5% and 96.1% accurate respectively. Each model revealed a unique combination of crystallographic attributes that affected twinning in the AZ31.

This thesis proposes a novel automated crack detection and characterisation method on unstructured surfaces using 3D point cloud. Crack detection on unstructured surfaces poses a challenge compared to flat surfaces such as pavements and concrete, which typically utilise image-based sensors. The detection method utilises a point cloud-based deep learning method to perform point-wise classification. The detected points are then automatically characterised to estimate the detected cracks’ properties such as width profile, orientation, and length. The proposed method enables the deployment of autonomous systems to conduct reliable surveys in environments risky to humans.

The advancements in robotic perception in the recent years have empowered robots to better execute tasks in various environments. The perception of objects in the robot work space significantly relies on how sensory data is represented. In this context, 3D models of object’s surfaces have been studied as a means to provide useful insights on shape of objects and ultimately enhance robotic perception. This involves several challenges, because sensory data generally presents artifacts, such as noise and incompleteness. To tackle this problem, we employ Gaussian Process Implicit Surface (GPIS), a non-parametric probabilistic reconstruction of object’s surfaces from 3D data points. This thesis investigates different configurations for GPIS, as a means to tackle the extraction of shape information. In our approach we interpret an object’s surface as the level-set of an underlying sparse Gaussian Process (GP) with variational formulation. Results show that the variational formulation for sparse GP enables a reliable approximation to the full GP solution. Experiments are performed on a synthetic and a real sensory data set. We evaluate results by assessing how close the reconstructed surfaces are to the ground-truth correspondences, and how well objects from different categories are clustered based on the obtained representation. Finally we conclude that the proposed solution derives adequate surface representations to reason about object shape and to discriminate objects based on shape information.
Framsteg inom robotperception de senaste åren har resulterat i robotar som är bättre på attutföra uppgifter i olika miljöer. Perception av objekt i robotens arbetsmiljö är beroende avhur sensorisk data representeras. I det här sammanhanget har 3D-modeller av objektytorstuderats för att ge användbar insikt om objektens form och i slutändan bättre robotperception. Detta innebär flera utmaningar, eftersom sensoriska data ofta innehåller artefakter, såsom brus och brist på data. För att hantera detta problem använder vi oss av Gaussian Process Implicit Surface (GPIS), som är en icke-parametrisk probabilistisk rekonstruktion av ett objekts yta utifrån 3D-punkter. Detta examensarbete undersöker olika konfigurationer av GPIS för att på detta sätt kunna extrahera forminformation. I vår metod tolkar vi ett objekts yta som nivåkurvor hos en underliggande gles variational Gaussian Process (GP) modell. Resultat visar att en gles variational GP möjliggör en tillförlitlig approximation av en komplett GP-lösningen. Experiment utförs på ett syntetisk och ett reellt sensorisk dataset. Vi utvärderar resultat genom att bedöma hur nära de rekonstruerade ytorna är till grundtruth- korrespondenser, och hur väl objektkategorier klustras utifrån den erhållna representationen. Slutligen konstaterar vi att den föreslagna lösningen leder till tillräckligt goda representationer av ytor för tolkning av objektens form och för att diskriminera objekt utifrån forminformation.

The flipped classroom has been increasingly used in higher education worldwide, and more recently in developing countries. The pedagogy involves a ‘flip’ of direct instruction being conducted online prior to class and learning activities demanding higher order thinking occurring in subsequent, face-to-face classrooms. While the flipped classroom has been well-researched in Western countries such as the USA, the UK and Australia, little is known about the implementation of the flipped classroom in a developing country like Vietnam. Here, the flipped classroom poses challenges to teachers’ and students’ traditional perspectives of teaching and learning, and to levels of infrastructure and training. To date, no studies have examined the perspectives of, and learning experiences in the flipped classroom for Vietnamese English as Foreign Language (EFL) undergraduate students. This study was conducted to address this gap. This study explored undergraduate students’ perspectives, and their learning experiences, in one case study university in Vietnam. The university had mandated the use of the flipped classroom in EFL courses in 2015 and the major aim of this study was to investigate how students were dealing with the pedagogy. Utilising symbolic interactionism as the theoretical perspective, the study employed two data collection methods, interviews, and observations. Semi-structured interviews were conducted with 20 EFL students and five EFL teachers; 30 observations of students’ learning activities occurred in both online learning and face-to-face classes. Data were thematically analysed to explore EFL students’ perspectives and learning experiences within a flipped classroom environment, and to triangulate these with the perspectives of the teachers responsible for carrying out the flipped classroom model. The study revealed five important findings. First, students showed their preferences for surface learning over deep learning in the flipped classroom. Second, higher-achieving students were engaged in deeper learning, but lower-achieving students struggled to move beyond surface learning. Third, students revealed limited understandings of the demands of flipped classroom learning; what was required to engage effectively and its strategic goals in EFL education. Fourth, students expressed a range of beliefs about self-regulated and metacognitive strategies, but these revealed inconsistencies across the cohort. Fifth, there were a range of individual and contextual factors that affected students’ surface learning in the flipped classroom. This study has several implications for Vietnamese higher education institutions wishing to adopt EFL flipped classroom learning. These include raising institutional awareness for preparing the management change agenda, focusing on students’ learning approaches and skills needed for the flipped classroom, and providing ongoing professional development and support for teachers and curriculum designers regarding theories underpinning the flipped classroom.

Exploring surfaces is an essential ability for humans, allowing them to interact with a large variety of objects within their environment. This ability to explore surfaces is also of a major interest in the development of a new generation of humanoid robots, which requires the development of more efficient artificial tactile sensing techniques. The details perceived by statically touching different surfaces of objects not only improve robotic hand performance in force-controlled grasping tasks but also enables the feeling of vibrations on touched surfaces. This thesis presents an extensive experimental study of object surface exploration using biologically-Inspired tactile-enabled robotic fingers. A new multi-modal tactile sensor, embedded in both versions of the robotic fingertips (similar to the human distal phalanx) is capable of measuring the heart rate with a mean absolute error of 1.47 bpm through static explorations of the human skin. A two-phalanx articulated robotic finger with a new miniaturized tactile sensor embedded into the fingertip was developed in order to detect and classify surface textures. This classification is performed by the dynamic exploration of touched object surfaces. Two types of movements were studied: one-dimensional (1D) and two-dimensional (2D) movements. The machine learning techniques - Support Vector Machine (SVM), Multilayer Perceptron (MLP), Random Forest, Extra Trees, and k-Nearest Neighbors (kNN) - were tested in order to find the most efficient one for the classification of the recovered textured surfaces. A 95% precision was achieved when using the Extra Trees technique for the classification of the 1D recovered texture patterns. Experimental results confirmed that the 2D textured surface exploration using a hemispheric tactile-enabled finger was superior to the 1D exploration. Three exploratory velocities were used for the 2D exploration: 30 mm/s, 35 mm/s, and 40 mm/s. The best classification accuracy of the 2D recovered texture patterns was 99.1% and 99.3%, using the SVM classifier, for the two lower exploratory velocities (30 mm/s and 35mm/s), respectively. For the 40 mm/s velocity, the Extra Trees classifier provided a classification accuracy of 99.4%. The results of the experimental research presented in this thesis could be suitable candidates for future development.

Wildlife ecotourism is becoming a well-established industry due to its ability to contribute to local economies and the growing tourist demand for opportunities to observe endangered or rare species. Wildlife ecotourism is also recognized for its ability to provide free choice-learning settings for visitors, through the use of environmental interpretation programs. The process of environmental interpretation is a communication phenomenon thought to hold the potential to contribute to conservation by educating and raising awareness amongst tourists about environmental issues. Using a qualitative phenomenological research design, this research examined the environmental interpretation programs of great white shark ecotourism operators in Gansbaai, South Africa, from the perspective of tourists. Findings indicated that while tourists did not primarily choose to embark on shark tourism excursions to learn more about the species, many participants became slightly more informed about great whites and the surrounding environment after their experience. It was also found that most participants did not experience nervousness or fear when in the water with great whites, but instead felt an emotional connection and appreciation for the animal, causing a shift towards pro-conservation attitudes.

Published Article
Work-Integrated Learning (WIL) is a form of Experiential Learning (EL) which implies learning by experience. This article represents the findings of a study regarding the roles of WIL and how such roles can be quantified when measured against the achievement of Critical Cross-Field Outcomes (CCFOs). The study was based on an empirical mixed-method triangulation, which allowed the researchers to use both qualitative and quantitative methods to address the research problem. The sample size is 35, constituting the third and fourth-year groups in the Hospitality Management programme at a higher education institution in South Africa. The results of the quantitative study indicate that the students have identified various roles for WIL whereas the quantitative investigation revealed that students are of the opinion that WIL contributes significantly towards the achievement of CCFOs. WIL therefore contributes to skills development in general and to the attainment of skills and attributes as represented by the CCFOs in particular. Recommendations regarding the implications of the study are made for curriculation purposes as well as for credit values to be attached to WIL.

The implied volatility surface plays an important role for Front office and Risk Management functions at Nasdaq and other financial institutions which require mark-to-market of derivative books intraday in order to properly value their instruments and measure risk in trading activities. Based on the aforementioned business needs, being able to calibrate an end of day implied volatility surface based on new market information is a sought after trait. In this thesis a statistical learning approach is used to calibrate the implied volatility surface intraday. This is done by using OMXS30-2019 implied volatility surface data in combination with market information from close to at the money options and feeding it into 3 Machine Learning models. The models, including Feed Forward Neural Network, Recurrent Neural Network and Gaussian Process, were compared based on optimal input and data preprocessing steps. When comparing the best Machine Learning model to the benchmark the performance was similar, indicating that the calibration approach did not offer much improvement. However the calibrated models had a slightly lower spread and average error compared to the benchmark indicating that there is potential of using Machine Learning to calibrate the implied volatility surface.
Implicita volatilitetsytor är ett viktigt vektyg för front office- och riskhanteringsfunktioner hos Nasdaq och andra finansiella institut som behöver omvärdera deras portföljer bestående av derivat under dagen men också för att mäta risk i handeln. Baserat på ovannämnda affärsbehov är det eftertraktat att kunna kalibrera de implicita volatilitets ytorna som skapas i slutet av dagen nästkommande dag baserat på ny marknadsinformation. I denna uppsats används statistisk inlärning för att kalibrera dessa ytor. Detta görs genom att uttnytja historiska ytor från optioner i OMXS30 under 2019 i kombination med optioner nära at the money för att träna 3 Maskininlärnings modeller. Modellerna inkluderar Feed Forward Neural Network, Recurrent Neural Network och Gaussian Process som vidare jämfördes baserat på data som var bearbetat på olika sätt. Den bästa Maskinlärnings modellen jämfördes med ett basvärde som bestod av att använda föregående dags yta där resultatet inte innebar någon större förbättring. Samtidigt hade modellen en lägre spridning samt genomsnittligt fel i jämförelse med basvärdet som indikerar att det finns potential att använda Maskininlärning för att kalibrera dessa ytor.

Surface mount technology is a common way to assembly electrical components onto PrintedCircuit Boards (PCB). To assemble the components, solder paste is used. One way to apply solderpaste onto PCB is jet printing.The quality of the solder paste deposits on the PCB depends on the properties of the solder pasteand the ejection parameters settings of the jet printer. Every solder paste is unique with its owncharacteristics. Solder paste dots are of good quality if the positioning of the dot is good, the dotis circular, and the number of satellites is at a minimum. A satellite is a droplet that has fallenoutside the main droplet. The parameters that have the most effect on the solder paste are thewaveform parameters Rise time and Voltage level.This master thesis examined the possibility to design and implement a feedback-based machinelearning algorithm that can find the most suitable value for the Rise time and Voltage level, thatgives good quality of the solder paste deposits. The algorithm that was used was a ReinforcementLearning algorithm. Reinforcement Learning is a reward-based learning algorithm where an agentlearns to interact with an environment by using trial and error. The specific algorithm that wasused was a Deep-Q-Learning algorithm. In this master thesis, it was also examined how the cameraresolution affects the decision of the algorithm. To see the implication of the camera resolution,two machines were used, an older and a newer machine were used where one of the biggestdifferences is that the camera resolution.It was concluded that a Deep-Q-Learning algorithm can be used to find the most suitable value forthe waveform parameters Rise time and Voltage level, which results in specified quality of thesolder paste deposits. It was also concluded that the algorithm converges faster for a lower cameraresolution, but the results obtained are more optional with the higher camera resolution.
Ytmontering är en metod som används för att montera elektriska komponenter på kretskort. Föratt kunna montera komponenterna används lödpasta. En teknik för att applicera lödpasta påkretskort är jet printing.Kvaliteten på lödpastavolymen på ett kretskort beror dels på egenskaperna hos lödpastan, dels påutskjutningssparametrarna hos jetprintern. Varje lödpasta är unik med hänsyn till flödesegenskaper. En lödpastadeposition har god kvalitet om depositionen har en bra position, omdepositionen är cirkulär och om mängden satelliter är minimal. En satellit är en droppe lödpastasom fallit utanför huvuddepositionen. Parametrarna som har störst effekt på lödpasta ärvågformsparameterna stigtid och spänningsnivå.Detta examensarbete undersökte möjligheten att hitta en feedbackbaserad maskininlärningsalgoritm som kan hitta de mest lämpliga värdena för stigtiden och spänningsnivå som ger godkvalitet på lödpastadepositionen. Algoritmen som användes var en Förstärkande inlärningsalgoritm.Förstärkande inlärning är en belöningsbaserad inlärningsalgoritm där en agent lär sig attinteragera med en miljö genom att använda trial and error. Den specifika algoritmen som användesvar en Deep-Q-Learning-algoritm. I examensarbetet undersöktes även hur kameraupplösningenspåverkar algoritmen och dess beslut. För att undersöka detta användes två maskiner, en nyare ochäldre version där att kameraupplösningen är lägre.Slutsatsen som drogs var att en Deep-Q-Learning-algoritm kan användas för att hitta det mestlämpliga värdena för vågformsparametrarna stigtid och spänningsnivå. En annan slutsats somdrogs var att algoritmen konvergerade snabbare när kameraupplösningen är lägre. Parapeternasom är optimala för den kameran med lägre upplösning är inte optimala för den kameran medhögre upplösning.

Uno dei problemi più importanti della 3D Computer Vision è il cosiddetto surface matching, che consiste nel trovare corrispondenze tra oggetti tridimensionali. Attualmente il problema viene affrontato calcolando delle feature locali e compatte, chiamate descrittori, che devono essere riconosciute e messe in corrispondenza al mutare della posa dell'oggetto nello spazio, e devono quindi essere invarianti rispetto all'orientazione. Il metodo più usato per ottenere questa proprietà consiste nell'utilizzare dei Local Reference Frame (LRF): sistemi di coordinate locali che forniscono un'orientazione canonica alle porzioni di oggetti 3D che vengono usate per calcolare i descrittori. In letteratura esistono diversi modi per calcolare gli LRF, ma fanno tutti uso di algoritmi progettati manualmente. Vi è anche una recente proposta che utilizza reti neurali, tuttavia queste vengono addestrate mediante feature specificamente progettate per lo scopo, il che non permette di sfruttare pienamente i benefici delle moderne strategie di end-to-end learning. Lo scopo di questo lavoro è utilizzare un approccio data-driven per far imparare a una rete neurale il calcolo di un Local Reference Frame a partire da point cloud grezze, producendo quindi il primo esempio di end-to-end learning applicato alla stima di LRF. Per farlo, sfruttiamo una recente innovazione chiamata Spherical Convolutional Neural Networks, le quali generano e processano segnali nello spazio SO(3) e sono quindi naturalmente adatte a rappresentare e stimare orientazioni e LRF. Confrontiamo le prestazioni ottenute con quelle di metodi esistenti su benchmark standard, ottenendo risultati promettenti.

The focus of this study was to characterize the surface water and groundwater hydrology of borrow-pit wetlands located within the borders of the Lewisville Lake Environmental Learning Area (LLELA), east of the Elm Fork of the Trinity River. The wetlands were excavated into alluvial deposits downstream of the Lewisville Lake Dam. Both surface water and groundwater contribute to the hydro-period of the borrow-pit wetlands. Nearby marshes exhibit characteristics of groundwater discharge. Salinity in groundwater-fed wetlands could affect establishment of vegetation, as suggested from plant surveys. Surface water input from storm events dilutes salinity levels. Management of LLELA wetlands should include long-term evaluation of hydrology and plantings to enhance habitat. Plans for additional wetlands should consider both surface water and groundwater inputs.

One purpose of this study was to identify the relationship, if any, that existed between method of scheduling and achievement in reading and mathematics. A second purpose of this study was to identify the relationship, if any, that existed between method of scheduling and student attitudes toward school and learning. A third purpose of the study was to identify what teachers using parallel block scheduling perceived as positive and negative aspects of parallel block scheduling.Data from ISTEP (Indiana Statewide Testing for Educational Progress) scores, student questionnaires, and teacher interviews were used to compare the parallel block scheduled school and the surface scheduled school. ISTEP scores were compared using One Way Analyses of Variance to check equivalency of the two schools at the beginning and at the end of the study and Repeated Measures Analyses of Variance to test the hypotheses. Student questionnairesmeasuring student attitudes toward school and learning were compared for the two schools using Repeated Measures Analyses of Variance. Teachers at the parallel block scheduled school were interviewed to analyze their perceptions of parallel block scheduling.Statistically significant differences were found in mathematics achievement in favor of the parallel block scheduled school. No statistically significant differences were found in reading achievement between the parallel block scheduled school and the surface scheduled school. Statistically significant differences in students' attitudes toward school and learning between the two types of scheduling were found in three out of the four categories. Differences were found in students' beliefs about how well they were learning, students' attitudes toward themselves as learners, and students' beliefs about how others see them as learners in favor of the parallel block scheduled school. No difference was found in students' attitudes toward school. Responses from interviews of teachers using parallel block scheduling indicated that the teachers felt parallel block scheduling had benefitted students in terms of achievement, attitudes toward school and learning, and in their effectiveness as teachers.
Department of Educational Leadership

Screen media activity such as watching TV and videos, playing video games, and using social media has become a popular leisure activity for children and adolescents. The effect of screen time has been a highly debated topic; however, there is still very little known about it. Using a dataset from the Adolescent Brain Cognitive Development longitudinal study 4 217 young adolescents, that met the requirements, could be retrieved for this thesis project after processing of the data. This thesis project investigated causal order between genetic effect (cognitive performance Polygenic scores (PGSs)), screen time activity, brain morphology (structural Magnetic Resonance Imaging (sMRI) for surface area and cortical thickness), lack of perseverance, and cognitive performance (crystallized IQ) with a machine learning algorithm DirectLiNGAM. A clear correlation between screen media activity and PGS was found for all types of screen time activities but only video games and social media correlated to the global surface area. Furthermore,  TV and video seem to affect lack of perseverance, and lack of perseverance, in turn, affects time spent on video games. These findings imply that different types of social media are not as alike as we thought and can affect adolescents differently. Taken together, these findings support previous research on screen media activity's effect on lack of perseverance, brain morphology, and cognitive performance, and propose new causal inference between genetics and screen time. Lastly, the algorithm used in this thesis project inferred reasonable causal orders and can be seen as a very good complement to today's causal modeling.
Skärmaktivitet som att titta på TV och video, spela videospel och använda sociala medier har blivit en populär fritidsaktivitet för barn och ungdomar. Effekten av skärmtid har varit ett mycket debatterat ämne; det finns dock fortfarande mycket lite kunskap om det. Med hjälp av datasetet från Adolescent Brain Cognitive Development långtidsstudien kunde 4 217 ungdomar, som uppfyllde specifika krav, väljas ut för detta avhandlingsprojekt efter bearbetning av datan. Detta avhandlingsprojekt undersökte kausal ordning mellan genetisk effekt (Polygenic scores (PGS) för kognitiv prestation), skärmtidsaktivitet, hjärnmorfologi (strukturell Magnet Resonans Imaging (sMRI) för hjärnans ytarea och hjärnbarks tjocklek), brist på ihärdighet och kognitiv förmåga (kristalliserad IQ) med en maskininlärningsalgoritm DirectLiNGAM. Tydlig korrelation mellan skärmaktivitet och PGS hittades för alla typer av skärmaktiviteter men endast videospel och sociala medier korrelerade till den globala ytarean. Dessutom verkar TV och video påverka brist på ihärdighet och brist på ihärdighet i sin tur påverkar hur mycket tid som spenderas på videospel. Dessa resultat antyder att olika typer av sociala medier inte är så lika som vi trodde och kan påverka ungdomar olika. Sammanlagt stöder dessa upptäckter tidigare forskning om skärmtidseffekt på brist på ihärdighet, hjärnmorfologi och kognitiv förmåga och föreslår en ny kausal inferens mellan genetik och skärmtid. Slutligen ledde algoritmen som användes i detta avhandlingsprojekt fram till rimliga kausala ordningar och kan ses som ett mycket bra komplement till dagens kausala modellering.

La surface du cortex cérébral est très convoluée, avec un grand nombre de plis, les sillons corticaux. Ces plis sont extrêmement variables d'un individu à l'autre. Cette grande variabilité constitue un problème pour de nombreuses applications en neurosciences et en imagerie cérébrale. Un problème central est que les sillons cérébraux ne sont pas la bonne unité pour décrire les plis sur la surface corticale. En particulier, leur géométrie (forme) et leur topologie (branches, nombre de pièces) sont très variables. Les "Plis de passages" (PPs) peuvent expliquer une partie de cette variabilité. Le concept de PPs a été introduit pour la première fois par Gratiolet (1854) pour décrire les gyri transversaux qui interconnectent les deux côtés d'un sillon, sont fréquemment enfouis dans la profondeur de ces sillons, et sont parfois apparents sur la surface corticale. En tant que caractéristique intéressante du processus de plissement cortical, la connectivité structurelle sous-jacente des PP a également suscité beaucoup d'intérêt.Cependant, la difficulté d'identifier les PPs et le manque de méthodes systématiques pour les détecter automatiquement ont limité leur utilisation. Cette thèse vise à détecter et à caractériser les PPs sur la surface corticale tant du point de vue de la morphologie que de la connectivité. Elle s'articule autour de deux axes de recherche principaux : 1. Définition d'un processus de détection des PPs basé sur l'apprentissage automatique et utilisant leurs caractéristiques géométriques (ou morphologiques).2. Étudier les relations entre les PP et leur connectivité structurelle sous-jacente, et poursuivre le développement de modèles d'apprentissage automatique multimodaux. Dans la première partie, nous présentons une méthode de détection automatique des PP sur le cortex en fonction des caractéristiques morphologiques locales proposées dans (Bodin et al., 2021). Pour enregistrer les caractéristiques morphologiques locales de chaque sommet de la surface corticale, nous avons utilisé la méthode de profilage de la surface corticale (Li et al., 2010). Ensuite, le problème de reconnaissance tridimensionnelle des PP est converti en un problème de classification d'image bidimensionnelle avec un déséquilibre de classe où plus de points dans le STS sont des non-PP que des PP. Pour résoudre ce cas, nous proposons un modèle “Ensemble SVM” (EnsSVM) avec une stratégie de rééquilibrage. Les résultats expérimentaux et les analyses statistiques quantitatives montrent l'efficacité et la robustesse de notre méthode. Dans la deuxième partie, nous étudions la connectivité structurelle, en particulier les fibres U à courte portée, qui sous-tend la localisation des PPs, et proposons une nouvelle approche pour étudier la densité des terminaisons des fibres U sur la surface corticale. Nous émettons l'hypothèse que les PPs sont situés dans des régions de haute densité de terminaisons de fibres U croisées. En effet, nos analyses statistiques montrent une corrélation de robustesse entre les PPs et la densité de terminaisons des fibres U. De plus, nous discutons de l'impact de l'hétérogénéité de la connectivité dans le STS sur les résultats de l'apprentissage automatique. Enfin, nous investiguons l'utilisation de cartes de myéline comme un complément à la connectivité structurelle
The surface of the cerebral cortex is very convoluted, with a large number of folds, the cortical sulci. Moreover, these folds are extremely variable from one individual to another. This great variability is a problem for many applications in neuroscience and brain imaging. One central problem is that cerebral sulci are not the good unit to describe folding over the cortical surface. In particular, their geometry (shape) and topology (branches, number of pieces) are very variable. “Plis de passages” (PPs) or “annectant gyri” can explain part of the variability. The concept of PPs was first introduced by Gratiolet (1854) to describe transverse gyri that interconnect both sides of a sulcus, are frequently buried in the depth of these sulci, and are sometimes apparent on the cortical surface. As an interesting feature of the cortical folding process, the underlying structural connectivity of PPs also generated a lot of interest. However, the difficulty of identifying PPs and the lack of systematic methods to automatically detecting them limited their use. This thesis aims to detect and characterise the PPs on the cortical surface from both morphology and connectivity aspects. It was structured around two main research axes: 1. Definition of a machine learning-based PPs detection process using their geometrical (or morphological) characteristics. 2. Investigate the relationships between PPs and their un- derlying structural connectivity, and further development of multi-modal machine learning models. In the first part, we present a method to detect the PPs on the cortex automatically according to the local morphological characteristics proposed in (Bodin et al., 2021), To record the local morphological patterns for each vertex on the cortical surface, we used the cortical surface profiling method (Li et al., 2010). After that, the three-dimensional PP recognition problem is converted to a two-dimensional image classification problem of class-imbalance where more points in the STS are non-PPs than PPs. To solve this case, we propose an ensemble SVM model (EnsSVM) with a rebalancing strategy. Experimental results and quantitative statistics analyses show the effectiveness and robustness of our method. In the second part, we study the structural connectivity, particularly short-range U-fibers, underlying the location of PPs, and propose a new approach to study the density of U-fiber terminations on the cortical surface. We hypothesize that the PPs are located in regions of high density of intercrossing U-fibers termination. Indeed, our statistical analyses show a robustness correlation between PPs and U-fibers termination density. Moreover, we discuss the impact of connectivity heterogeneity in the STS on the machine learning results, and the myelin map is then used as a supplement to the structural connectivity

La creatividad es una habilidad de especial interés para el desarrollo humano dado que es una de las dimensiones que permite al individuo y en última instancia a la sociedad enfrentarse a nuevos problemas y retos de forma satisfactoria. Además de entender la creatividad como una serie de factores relativos al individuo creativo, debe tenerse en cuenta que el grado de motivación intrínseca, el entorno y otros factores sociales pueden tener un efecto relevante sobre el desarrollo de esta importante habilidad, por lo que resulta de interés explorarla en el contexto de utilización de tecnologías de la información. En particular, dado que los procesos comunicativos, el intercambio de ideas y la interacción colaborativa entre individuos son un pilar fundamental en los procesos creativos, y también que en gran medida todas ellas son características mayormente facilitadas por las mesas interactivas, una de las principales contribuciones de esta tesis consiste precisamente en la exploración de la idoneidad de las superficies interactivas en tareas creativas colaborativas de construcción en estudiantes adolescentes. Partiendo del estudio realizado, que aporta evidencia empírica acerca de la adecuación de las superficies interactivas como tecnología de potencial para el fomento de la creatividad, esta tesis presenta AGORAS: un middleware para la construcción de ecosistemas de juegos 2D para mesas interactivas, y cuya idea final es entender actividades de aprendizaje más enriquecedoras como aquellas que permiten la propia creación de juegos y su posterior consumo. En el contexto de esta tesis también se ha desarrollado un toolkit básico para construcción de interfaces de usuario para superficies interactivas, se ha desarrollado un modelo de ecosistema basado en entidades que son simulables de acuerdo a leyes físicas; y se ha dotado al modelo de aproximación basada en reglas de comportamiento enriquecidas con expresiones dataflows y de su correspondiente editor para superficies.
Catalá Bolós, A. (2012). AGORAS: Augmented Generation of Reactive Ambients on Surfaces. Towards educational places for action, discussion and reflection to support creative learning on interactive surfaces [Tesis doctoral no publicada]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/16695
Palancia