Made available in DSpace on 2016-06-02T20:19:19Z (GMT). No. of bitstreams: 1 4543.pdf: 1379957 bytes, checksum: 76a64c72298bfebb3102cd24d039fed8 (MD5) Previous issue date: 2012-07-18 / Universidade Federal de Minas Gerais / Apesar do uso generalizado do Incremental Shuttle Walk Test (ISWT), não há estudos anteriores que avaliaram este teste na população de mulheres obesas. Portanto, o objetivo principal foi testar a reprodutibilidade e confiabilidade deste teste e sua validade em comparação com um teste em esteira (TECP). Quarenta e seis mulheres realizaram três ISWT em um corredor de 10 m, dois no mesmo dia e o terceiro com intervalo de dois a sete dias e um TECP. A idade e o índice de massa corporal médios foram de 32 anos e 35 kg/m², respectivamente. Houve diferença significativa entre o primeiro e o segundo ISWT para as variáveis cardiovasculares e distância percorrida (ISWD) (frequência cardíaca=139±16 vs. 150±18 bpm, p=0,004; pressão arterial sistólica=157±9 vs. 159±19 mmHg, p<0,001; ISWD=413±88 vs. 463±92 m,p=0,009). Não houve diferença significante para nenhuma variável mensurada entre os valores obtidos no pico do segundo com o terceiro teste, além disso, o segundo ISWT apresentou concordância com o terceiro quando analisada a diferença entre as médias das diferenças. Houve boa e excelente confiabilidade entre o segundo e terceiro ISWT quando analisada a distância (ISWD) (ICC=0,93), ventilação (ICC=0,78), consumo de oxigênio (VO2) (ICC=0,90) e produção de dióxido de carbono (ICC=0,86). O VO2 pico obtido no TECP foi significativamente correlacionado com a ISWD (r=0,54, p<0,05) e com o VO2 pico no ISWT (r=0,64, p<0,05). Em conclusão, nossos achados sugerem que o ISWT pode ser uma ferramenta válida, confiável e reprodutível para avaliação da capacidade física de mulheres adultas com obesidade.

abstract: The dynamic nature of Navajo or Diné culture is continuing to be constrained by a mechanistic planning paradigm supporting delivery of colonial subdivisions across the land. Poor housing and subdivision conditions levy pressures on the Navajo People that reduce their ability to cope with environmental, financial and social pressures. This study has taken this complex social justice related health challenge to heart through a 2015-2016 school year of Arizona State University dissertation driven, community-based participatory action research with high school students from Navajo Preparatory School (NPS) in Farmington, New Mexico and community participants from the Shiprock Chapter of the Navajo Nation. Fieldwork focused on case study analysis of cluster settlements across the Navajo Northern Agency and existing subdivisions within the town of Shiprock to develop the Framework for a transformational Navajo model of the Pattern Language (Alexander et al, 1977) for new neighborhood design. Pattern data supporting the Framework was generated at the linked scales of the Navajo nuclear Family Camp, the extended family Cluster Camp, and the community-scaled Constellation Settlement “spatial model” that is proposed by this study as new neighborhood planning model. An ethnographic research methodology was employed with students, faculty, Board leadership and neighboring Shiprock Chapter and Shiprock Planning Commission research participants. The study’s research methodology was anchored by a pioneering Indigenous Planning high school course that was housed within the School’s International Baccalaureate curriculum. Goals for student education in Indigenous Planning theory and much needed Diné planning-based language building were married with practical aims for use of the Diné Pattern Language and Constellation Settlement spatial model for anticipated Shiprock Chapter housing projects. / Dissertation/Thesis / Doctoral Dissertation Design 2017

Atualmente diversos sistemas sensoriais fornecem dados em fluxos e essas observações medidas são frequentemente de alta dimensionalidade, ou seja, o número de variáveis medidas é grande, e as observações chegam em sequência. Este é, em particular, o caso de sistemas de visão em robôs. Aprendizagem supervisionada e não-supervisionada com esses fluxos de dados é um desafio, porque o algoritmo deve ser capaz de aprender com cada observação e depois descartá-la antes de considerar a próxima, mas diversos métodos requerem todo o conjunto de dados a fim de estimar seus parâmetros e, portanto, não são adequados para aprendizagem em tempo real. Além disso, muitas abordagens sofrem com a denominada maldição da dimensionalidade (BELLMAN, 1961) e não conseguem lidar com dados de entrada de alta dimensionalidade. Para superar os problemas descritos anteriormente, este trabalho propõe um novo modelo de rede neural probabilístico e incremental, denominado Local Projection Incremental Gaussian Mixture Network (LP-IGMN), que é capaz de realizar aprendizagem perpétua com dados de alta dimensionalidade, ou seja, ele pode aprender continuamente considerando a estabilidade dos parâmetros do modelo atual e automaticamente ajustar sua topologia levando em conta a fronteira do subespaço encontrado por cada neurônio oculto. O método proposto pode encontrar o subespaço intrísico onde os dados se localizam, o qual é denominado de subespaço principal. Ortogonal ao subespaço principal, existem as dimensões que são ruidosas ou que carregam pouca informação, ou seja, com pouca variância, e elas são descritas por um único parâmetro estimado. Portanto, LP-IGMN é robusta a diferentes fontes de dados e pode lidar com grande número de variáveis ruidosas e/ou irrelevantes nos dados medidos. Para avaliar a LP-IGMN nós realizamos diversos experimentos usando conjunto de dados simulados e reais. Demonstramos ainda diversas aplicações do nosso método em tarefas de reconhecimento de imagens. Os resultados mostraram que o desempenho da LP-IGMN é competitivo, e geralmente superior, com outras abordagens do estado da arte, e que ela pode ser utilizada com sucesso em aplicações que requerem aprendizagem perpétua em espaços de alta dimensionalidade. / Nowadays several sensory systems provide data in ows and these measured observations are frequently high-dimensional, i.e., the number of measured variables is large, and the observations are arriving in a sequence. This is in particular the case of robot vision systems. Unsupervised and supervised learning with such data streams is challenging, because the algorithm should be capable of learning from each observation and then discard it before considering the next one, but several methods require the whole dataset in order to estimate their parameters and, therefore, are not suitable for online learning. Furthermore, many approaches su er with the so called curse of dimensionality (BELLMAN, 1961) and can not handle high-dimensional input data. To overcome the problems described above, this work proposes a new probabilistic and incremental neural network model, called Local Projection Incremental Gaussian Mixture Network (LP-IGMN), which is capable to perform life-long learning with high-dimensional data, i.e., it can continuously learn considering the stability of the current model's parameters and automatically adjust its topology taking into account the subspace's boundary found by each hidden neuron. The proposed method can nd the intrinsic subspace where the data lie, which is called the principal subspace. Orthogonal to the principal subspace, there are the dimensions that are noisy or carry little information, i.e., with small variance, and they are described by a single estimated parameter. Therefore, LP-IGMN is robust to di erent sources of data and can deal with large number of noise and/or irrelevant variables in the measured data. To evaluate LP-IGMN we conducted several experiments using simulated and real datasets. We also demonstrated several applications of our method in image recognition tasks. The results have shown that the LP-IGMN performance is competitive, and usually superior, with other stateof- the-art approaches, and it can be successfully used in applications that require life-long learning in high-dimensional spaces.

Le développement durable est la conceptualisation de la transition imposée de la Société humaine vers un modede développement soutenable pour la planète, pour trouver une solution de sortie à la crise de l’Environnement. Acet égard, l’écoconception est l’une des solutions que le monde industriel et entrepreneurial se propose de mettreen application. Elle consiste à prendre en compte les impacts environnementaux sur l’ensemble du cycle de vied’un produit (bien ou service) lors de la conception de celui-ci. L’intégration de la dimension Environnementsoulève une triple complexité : (i) celle liée à la nature multicritère de la dimension Environnement, (ii) celle liéeà la compréhension de la dimension Environnement par les acteurs de la conception de produits dont la résistancepeut être un frein à une bonne intégration, et (iii) celle liée au processus de conception et d’innovation d’un produit,spécialement dans le cas des produits complexes. Or malgré la multitude d’outils d’écoconception qui a étédéveloppée, peu de sujets de recherche s’intéressent à la prise en compte à la fois des aspects techniques etorganisationnels, lors de l’intégration de la dimension Environnement en phase d’innovation d’un produitcomplexe. Cette thèse vise donc à expérimenter, chez un constructeur d’automobiles, une stratégie d’intégrationde l’évaluation environnementale d’innovations incrémentales nommée E3PICS (Methodology of an Evolutiveintegration of the Evaluation of the Environmental Performances of Innovative Complex Sub-systems). Mise enforme par des contraintes à la fois techniques et organisationnelles, la stratégie E3PICS emploie une démarcheprogressive d’intégration d’un référentiel évolutif d’écoconception dans les processus d’ingénierie avancéed’éléments innovants qui seront raccordés à des projets de développement des véhicules. La première étape estcelle de la conception itérative du référentiel évolutif d’écoconception avec l’équipe écoconception (au sein duservice Environnement) et les pilotes d’innovations, permettant ainsi un apprentissage croisé. La deuxième étapeest celle de l’accompagnement de l’ensemble des pilotes d’innovations dans l’utilisation systématique duréférentiel d’écoconception. La troisième étape consiste à développer un outil analytique d’évaluation de l’impactd’innovations sur la recyclabilité en fin de vie des véhicules. Elle nécessite la création de modèles de véhiculesafin de contourner le manque de données sur le système complet en cours de conception et de faire une projectioncet impact. La quatrième et dernière étape concerne le développement d’un deuxième outil analytique pourl’évaluation de l’impact d’innovations sur la performance environnementale sur le cycle de vie des véhicules. Dela même manière, cet outil nécessite une méthode de développement de modèles environnementaux de véhicules ;cependant, pour les impacts environnementaux calculés sur le cycle de vie, contrairement à la recyclabilité qui estcalculée sur le véhicule en fin de vie, ces modèles sont basés sur le traitement par classification ascendantehiérarchique de résultats d’analyse de cycle de vie de véhicules. Les expérimentations ont été réalisées chez PSAPeugeot Citroën. La stratégie E3PICS a permis d’intégrer dans les processus d’innovation, l’utilisation duréférentiel d’écoconception et de systématiser son utilisation, dans l’optique d’une amélioration continue pérennedes véhicules du constructeur. / Sustainable development is the conceptualization of transition imposed from human society towards a sustainableway of development for the world to find a solution to the crisis of the Environment. In this regard, ecodesign isone of the solutions that the industrial and business world proposes to implement. It consists of taking into accountthe environmental impacts throughout the life cycle of a product (good or service) in the design of it. Integratingthe Environment dimension raises a triple complexity: (i) the one related to the multi-criteria nature of theEnvironment dimension, (ii) the one related to understanding the Environment dimension by actors designingproducts who may resist and be a barrier to successful integration, and (iii) the one related to process design andinnovation of a product, especially in the case of complex products. But despite the multitude of ecodesign toolsthat has been developed, few research topics are interested in taking into account both technical and organizationalaspects, while integrating environment in the innovation phase of a complex product. Therefore this thesis aims toexperiment, by a car manufacturer, an integration strategy of the environmental evaluation of incrementalinnovations, named E3PICS (Methodology of an Evolutive integration of the Evaluation of the EnvironmentalPerformances of Innovative Complex Sub-systems). Framed by constraints on the technical and organizationalconstraints, E3PICS strategy employs a progressive approach to integrate a scalable ecodesign repository inadvanced process engineering of innovative features that are connected to development projects vehicles. The firststep is the iterative design of the scalable ecodesign repository with the ecodesign team (in the EnvironmentDepartment) and innovation leaders, allowing cross learning. The second step is to accompany all the innovationleaders in the systematic use of the ecodesign repository. The third step is to develop an analytical tool for assessingthe impact of innovations on the recycling end of life vehicles. It requires the creation of models of vehicles tobypass the lack of data on the complete system under design and project impact. The fourth and final step is thedevelopment of a second analytical tool for assessing the impact of innovations on environmental performanceover the life cycle of vehicles. Similarly, the tool requires a method of developing environmental models ofvehicles; however, for environmental impacts calculated on the life cycle, unlike recyclability which is calculatedon the vehicle end of life, these models are based on the treatment by hierarchical clustering of vehicle life cycleassessment results . The experiments were performed at PSA Peugeot Citroen. The E3PICS strategy has helped tointegrate the processes of innovation using the eco-design repository and systematize its use in the context of asustainable continuous improvement of vehicle manufacturer.

O recente campo de Deep Learning introduziu a área de Aprendizagem de Máquina novos métodos baseados em representações distribuídas e abstratas dos dados de treinamento ao longo de estruturas hierárquicas. A organização hierárquica de camadas permite que esses métodos guardem informações distribuídas sobre os sinais sensoriais e criem conceitos com diferentes níveis de abstração para representar os dados de entrada. Este trabalho investiga o impacto de uma estrutura hierárquica inspirada pelas ideias apresentadas em Deep Learning, e com base na Incremental Gaussian Mixture Network (IGMN), uma rede neural probabilística com aprendizagem online e incremental, especialmente adequada para as tarefas de robótica. Como resultado, foi desenvolvida uma arquitetura hierárquica, denominada Hierarchical Incremental Gaussian Mixture Network (HIGMN), que combina dois níveis de IGMNs. As camadas de primeiro nível da HIGMN são capazes de aprender conceitos a partir de dados de diferentes domínios que são então relacionados na camada de segundo nível. O modelo proposto foi comparado com a IGMN em tarefas de robótica, em especial, na tarefa de aprender e reproduzir um comportamento de seguir paredes, com base em uma abordagem de Aprendizado por Demonstração. Os experimentos mostraram como a HIGMN pode executar três diferentes tarefas em paralelo (aprendizagem de conceitos, segmentação de comportamento, e aprendizagem e reprodução de comportamentos) e sua capacidade de aprender um comportamento de seguir paredes e reproduzi-lo em ambientes desconhecidos com novas informações sensoriais. A HIGMN conseguiu reproduzir o comportamento de seguir paredes depois de uma única, simples e curta demonstração do comportamento. Além disso, ela adquiriu conhecimento de diferentes tipos: informações sobre o ambiente, a cinemática do robô, e o comportamento alvo. / The recent field of Deep Learning has introduced to Machine Learning new meth- ods based on distributed abstract representations of the training data throughout hierarchical structures. The hierarchical organization of layers allows these meth- ods to store distributed information on sensory signals and to create concepts with different abstraction levels to represent the input data. This work investigates the impact of a hierarchical structure inspired by ideas on Deep Learning and based on the Incremental Gaussian Mixture Network (IGMN), a probabilistic neural network with an on-line and incremental learning, specially suitable for robotic tasks. As a result, a hierarchical architecture, called Hierarchical Incremental Gaussian Mixture Network (HIGMN), was developed, which combines two levels of IGMNs. The HIGMN first-level layers are able to learn concepts from data of different domains that are then related in the second-level layer. The proposed model was compared with the IGMN regarding robotic tasks, in special, the task of learning and repro- ducing a wall-following behavior, based on a Learning from Demonstration (LfD) approach. The experiments showed how the HIGMN can perform parallely three different tasks concept learning, behavior segmentation, and learning and repro- ducing behaviors and its ability to learn a wall-following behavior and to perform it in unknown environments with new sensory information. HIGMN could reproduce the wall-following behavior after a single, simple, and short demonstration of the behavior. Moreover, it acquired different types of knowledge: information on the environment, the robot kinematics, and the target behavior.

Atualmente diversos sistemas sensoriais fornecem dados em fluxos e essas observações medidas são frequentemente de alta dimensionalidade, ou seja, o número de variáveis medidas é grande, e as observações chegam em sequência. Este é, em particular, o caso de sistemas de visão em robôs. Aprendizagem supervisionada e não-supervisionada com esses fluxos de dados é um desafio, porque o algoritmo deve ser capaz de aprender com cada observação e depois descartá-la antes de considerar a próxima, mas diversos métodos requerem todo o conjunto de dados a fim de estimar seus parâmetros e, portanto, não são adequados para aprendizagem em tempo real. Além disso, muitas abordagens sofrem com a denominada maldição da dimensionalidade (BELLMAN, 1961) e não conseguem lidar com dados de entrada de alta dimensionalidade. Para superar os problemas descritos anteriormente, este trabalho propõe um novo modelo de rede neural probabilístico e incremental, denominado Local Projection Incremental Gaussian Mixture Network (LP-IGMN), que é capaz de realizar aprendizagem perpétua com dados de alta dimensionalidade, ou seja, ele pode aprender continuamente considerando a estabilidade dos parâmetros do modelo atual e automaticamente ajustar sua topologia levando em conta a fronteira do subespaço encontrado por cada neurônio oculto. O método proposto pode encontrar o subespaço intrísico onde os dados se localizam, o qual é denominado de subespaço principal. Ortogonal ao subespaço principal, existem as dimensões que são ruidosas ou que carregam pouca informação, ou seja, com pouca variância, e elas são descritas por um único parâmetro estimado. Portanto, LP-IGMN é robusta a diferentes fontes de dados e pode lidar com grande número de variáveis ruidosas e/ou irrelevantes nos dados medidos. Para avaliar a LP-IGMN nós realizamos diversos experimentos usando conjunto de dados simulados e reais. Demonstramos ainda diversas aplicações do nosso método em tarefas de reconhecimento de imagens. Os resultados mostraram que o desempenho da LP-IGMN é competitivo, e geralmente superior, com outras abordagens do estado da arte, e que ela pode ser utilizada com sucesso em aplicações que requerem aprendizagem perpétua em espaços de alta dimensionalidade. / Nowadays several sensory systems provide data in ows and these measured observations are frequently high-dimensional, i.e., the number of measured variables is large, and the observations are arriving in a sequence. This is in particular the case of robot vision systems. Unsupervised and supervised learning with such data streams is challenging, because the algorithm should be capable of learning from each observation and then discard it before considering the next one, but several methods require the whole dataset in order to estimate their parameters and, therefore, are not suitable for online learning. Furthermore, many approaches su er with the so called curse of dimensionality (BELLMAN, 1961) and can not handle high-dimensional input data. To overcome the problems described above, this work proposes a new probabilistic and incremental neural network model, called Local Projection Incremental Gaussian Mixture Network (LP-IGMN), which is capable to perform life-long learning with high-dimensional data, i.e., it can continuously learn considering the stability of the current model's parameters and automatically adjust its topology taking into account the subspace's boundary found by each hidden neuron. The proposed method can nd the intrinsic subspace where the data lie, which is called the principal subspace. Orthogonal to the principal subspace, there are the dimensions that are noisy or carry little information, i.e., with small variance, and they are described by a single estimated parameter. Therefore, LP-IGMN is robust to di erent sources of data and can deal with large number of noise and/or irrelevant variables in the measured data. To evaluate LP-IGMN we conducted several experiments using simulated and real datasets. We also demonstrated several applications of our method in image recognition tasks. The results have shown that the LP-IGMN performance is competitive, and usually superior, with other stateof- the-art approaches, and it can be successfully used in applications that require life-long learning in high-dimensional spaces.

A estimativa de idade pela contagem das linhas incrementais de cemento (LC) adicionadas à idade média de erupção do dente analisado é um método tido como preciso e confiável por alguns autores, enquanto outros o rejeitam afirmando não haver forte correlação entre idade real e estimada. O objetivo do estudo foi avaliar a técnica descrita e verificar se há influência de patologias bucais na estimativa de idade, analisando-se, além do número de LC, correlação entre espessura de cemento e idade real. Foram preparadas por desgaste 31 lâminas transversais, de aproximadamente 30 m, de 25 dentes recém extraídos. As lâminas foram observadas, fotografadas e medidas em microscopia óptica. As LC das imagens foram realçadas com uso do software Image J 1.43s e as contagens foram feitas por um observador e dois observadores-controle. Houve correlação moderada de 0.58 para toda a amostra, com erro médio de 9,7 anos. Para dentes com alterações periodontais, a correlação foi de 0.03 e erro médio de 22,6 anos. Para dentes sem alterações periodontais, a correlação foi de 0.74 e erro médio de 1,6 anos. A espessura de cemento teve correlação com idade real de 0.69 para toda amostra, 0.25 para dentes com problemas periodontais e 0.75 para dentes sem problemas periodontais. A técnica das LC associada à medição de espessura de cemento mostrou-se confiável para dentes sem patologias periodontais, porém em dentes com patologias periodontais ou histórico/quadro clínico desconhecido, recomenda-se a realização de exames macroscópicos conjuntos para comparação. / Age estimation by counting incremental lines in cementum added to the average age of tooth eruption is considered an accurate and reliable method by some authors, while others reject it stating no strong correlation between estimated and actual age. The aim of this study was to evaluate this technique and check the influence of oral conditions on age estimation by analyzing both the number of cementum lines as well as the correlation between cementum thickness and actual age, on diseased teeth. Thirty one undecalcified ground cross sections of approximately 30 m, from 25 freshly extracted teeth were prepared, observed, photographed and measured. The images were enhanced with the use of software and the counts were made by one observer and two control-observers. There was moderate correlation ((r)=0.58) for the entire sample, with mean error of 9.7 years. For teeth with periodontal pathologies, the correlation was 0.03 with a mean error of 22.6 years. For teeth without periodontal pathologies, the correlation was 0.74 with mean error of 1.6 years. There was correlation of 0.69 between cementum thickness and actual age for the entire sample, 0.25 for teeth with periodontal problems and 0.75 for teeth without periodontal pathologies. The cementum lines technique associated with the measurement of cementum thickness was reliable for teeth without periodontal pathologies, but in periodontally diseased teeth or teeth with unknown history/clinical background, parallel macroscopic examinations should be conducted.

Les systèmes de dialogue incrémentaux sont capables d’entamer le traitement des paroles de l’utilisateur au moment même où il les prononce (sans attendre de signal de fin de phrase tel un long silence par exemple). Ils peuvent ainsi prendre la parole à n’importe quel moment et l’utilisateur peut faire de même (et interrompre le système). De ce fait, ces systèmes permettent d’effectuer une plus large palette de comportements de prise de parole en comparaison avec les systèmes de dialogue traditionnels. Cette thèse s’articule autour de la problématique suivante : est-il possible pour un système de dialogue incrémental d’apprendre une stratégie optimale de prise de parole de façon autonome? Tout d’abord, une analyse des mécanismes sous-jacents à la dynamique de prise de parole dans une conversation homme-homme a permis d’établir une taxonomie de ces phénomènes. Ensuite, une nouvelle architecture permettant de doter les systèmes de dialogues conventionnels de capacités de traitement incrémentales de la parole, à moindre coût, a été proposée. Dans un premier temps, un simulateur de dialogue destiné à répliquer les comportements incrémentaux de l’utilisateur et de la reconnaissance vocale a été développé puis utilisé pour effectuer les premier tests de stratégies de dialogue incrémentales. Ces dernières ont été développées à base de règles issues de l’analyse effectuée lors de l’établissement de la taxonomie des phénomènes de prise de parole. Les résultats de la simulation montrent que le caractère incrémental permet d’obtenir des interactions plus efficaces. La meilleure stratégie à base de règles a été retenue comme référence pour la suite. Dans un second temps, une stratégie basée sur l’apprentissage par renforcement a été implémentée. Elle est capable d’apprendre à optimiser ses décisions de prise de parole de façon totalement autonome étant donnée une fonction de récompense. Une première comparaison, en simulation, a montré que cette stratégie engendre des résultats encore meilleurs par rapport à la stratégie à base de règles. En guise de validation, une expérience avec des utilisateurs réels a été menée (interactions avec une maison intelligente). Une amélioration significative du taux de complétion de tâche a été constatée dans le cas de la stratégie apprise par renforcement et ce, sans dégradation de l’appréciation globale par les utilisateurs de la qualité du dialogue (en réalité, une légère amélioration a été constatée). / Incremental dialogue systems are able to process the user’s speech as it is spoken (without waiting for the end of a sentence before starting to process it). This makes them able to take the floor whenever they decide to (the user can also speak whenever she wants, even if the system is still holding the floor). As a consequence, they are able to perform a richer set of turn-taking behaviours compared to traditional systems. Several contributions are described in this thesis with the aim of showing that dialogue systems’ turn-taking capabilities can be automatically improved from data. First, human-human dialogue is analysed and a new taxonomy of turn-taking phenomena in human conversation is established. Based on this work, the different phenomena are analysed and some of them are selected for replication in a human-machine context (the ones that are more likely to improve a dialogue system’s efficiency). Then, a new architecture for incremental dialogue systems is introduced with the aim of transforming a traditional dialogue system into an incremental one at a low cost (also separating the turn-taking manager from the dialogue manager). To be able to perform the first tests, a simulated environment has been designed and implemented. It is able to replicate user and ASR behaviour that are specific to incremental processing, unlike existing simulators. Combined together, these contributions led to the establishement of a rule-based incremental dialogue strategy that is shown to improve the dialogue efficiency in a task-oriented situation and in simulation. A new reinforcement learning strategy has also been proposed. It is able to autonomously learn optimal turn-taking behavious throughout the interactions. The simulated environment has been used for training and for a first evaluation, where the new data-driven strategy is shown to outperform both the non-incremental and rule-based incremental strategies. In order to validate these results in real dialogue conditions, a prototype through which the users can interact in order to control their smart home has been developed. At the beginning of each interaction, the turn-taking strategy is randomly chosen among the non-incremental, the rule-based incremental and the reinforcement learning strategy (learned in simulation). A corpus of 206 dialogues has been collected. The results show that the reinforcement learning strategy significantly improves the dialogue efficiency without hurting the user experience (slightly improving it, in fact).

The quality of a product depends on the quality of the underlying process is a well known fact. Software development organizations have been struggling to decrease their cost, increase their ROI, reduce time-to-market, and enhance the quality of their products. This all depends upon the improvement in the processes they are following inside their organizations. A number of software process improvement models exist in market, e.g., CMMI, SPICE and Automotive SPICE. But before an organization can improve its development and management processes, it is very important to know whether it is following the right processes. There exist a number of software development process models, mainly categorized into Traditional and Agile, which provide the step-by-step guidance to develop and manage the software projects.The current thesis presents a study of software process improvement models in automotive industry, their weaknesses and strengths and presents a comparison of how do they relate to each other. This thesis also explores some software development models which are more famous in automotive industry, and the applicability of process improvement models in conjunction with the Agile software development models. A case study was performed at an automotive software supplier organization to investigate the experience of combining Agile practices with organization’s company-tailored software development model that was incorporating Automotive SPICE standards.

No description available.