Teses de Doutoramento
URI permanente para esta coleção:
Navegar
Entradas recentes
- Creativity support tool for sustainability: an LLM-Assisted Creative ProcessPublication . Rodrigues, Ana Isabel Mendonça; Campos, Pedro Filipe Pereira; Cabral, Diogo Nuno Crespo RibeiroPromoting new ideas is fundamental and promising for fostering innovation in various sectors. Creativity is the fundamental basis of innovation, serving as a springboard for the development of new and useful ideas in their application context, driving innovation in various spheres of society, from organizational processes to cultural transformations. Driven by the impact and evolution of Human-Computer Interaction (HCI) and the rise of the creative process through technology, this doctoral thesis focuses on user experience (UX) during the creative process through the development of a creativity support tool - System Approach to Idea Evaluation and Selection of Ecosystem (SAIESE) - based on a Large Language Model (LLM). The central question of this thesis is to analyze which functionalities can be implemented in creativity support tools, with the aim of increasing support for the creative process (i.e., generation, evaluation, and selection of ideas) and, consequently, improving the creative experience during the resolution of sustainable tasks or problems. In this context, the creative process was guided by three sustainability metrics (i.e., environmental impact, economic benefits, and social effects), ensuring that creativity and the process were directed to the application context of this research. The methodology adopted combines a hybrid research approach (qualitative and quantitative data) and a systematic literature review (2012-2025), which established the theoretical basis for the practical development of the prototype. The prototypes developed were accompanied by empirical validations guided by the research questions initially defined and by some research hypotheses. After extensive evaluations, SAIESE proved to be a successful example of a creativity support tool based on a large-scale language model, providing efficient and effective support throughout the creative process. The research concludes with the development of a conceptual approach that synthesizes the different dimensions of the research. We present the theoretical framework (Divergent Convergent Thinking Framework (DCTF)) that represents the spatialization of the creative process through the Ideas Space and Solution Space present in the Double Diamond Model. Next, we revisit the classic principles of CSTs in light of contemporary LLM-based technologies. Finally, we explore the dynamics of control, transparency, and trust in human-AI collaboration, particularly in the context of convergent thinking and AI-assisted decision making. In conclusion, this thesis proposes a new creativity support tool that combines human cognitive processes with the capabilities of an LLM to inspire future researchers, such as designers, who want to develop tools tailored to application contexts.
- Studies with galacto-oligosaccharides and lactic acid bacteria for the valorization of food by-productsPublication . Martins, Gonçalo Nuno Gouveia; Castilho, Paula Cristina Machado Ferreira; Gómez-Zavaglia, AndreaFood wastes and by-products’ generation raises humanitarian, economic and environmental concerns. The UN’s 12th Sustainable Development Goal promotes waste reduction and co-products’ valorisation along the food production chain. Prebiotic sugars galacto-oligosaccharides (GOS) resist digestion in the upper gastrointestinal tract, being metabolized by beneficial gut bacteria, supporting their proliferation, and promoting consumers’ health. GOS show cryoprotective potential towards lactic acid bacteria during freezing, freeze-drying, and storage, by replacing water molecules and forming a glass like structure around the bacteria, preventing cell damage. α-GOS can be obtained from natural sources and β-GOS by enzymatic synthesis from lactose. Removal of glucose formed during the synthesis increases the mixture’s health benefits. For the valorisation of food by-products, α-GOS from chickpeas’ and lentils’ cooking wastewaters were used for growing and the stabilization of two food-grade Lactiplantibacillus plantarum strains, and β-GOS were produced by two β-galactosidases immobilized in halloysite nanotubes and purified by fermentation with surplus yeast from the Madeiran brewing industry. Chickpeas’ yielded the most α-GOS, while lentils’ water contained more GOS of higher degree of polymerization and fewer simple sugars. L. plantarum CIDCA 83114 grew in cooking water-containing media, similarly to the standard microbiological media that uses glucose as carbon source. After freezing, freeze-drying, and storage for 3 weeks at 37 °C, GOS wastewaters were the most successful cryoprotectants towards L. plantarum WCFS1 strain, outperforming reference materials (sucrose and fructo oligosaccharides). After the enzymatic synthesis and purification by fermentation in a repeated batch operation, both mixtures’ final composition consisted in 41 % β-GOS, with unreacted lactose and galactose present, but only 1 % glucose. Food industry’s by-products are valuable sources of bioactive compounds and materials. Legumes’ cooking waters support the growth and protection of food-grade bacteria, while surplus yeast can be used in β-GOS’ purification. These added-value products can circle back to the food industry, tackling waste management and environmental concerns, while improving consumers’ health by the production of prebiotics, and probiotics with increased shelf-life.
- Emotional regulation assessment via multi-biosignal processing in a VR environment for neurorehabilitationPublication . Lima, Rodrigo Olival; Bermúdez i Badia, Sergi; Gamboa, Hugo Filipe Silveira; Cameirão, MÛnica da SilvaEmerging immersive technologies and physiological computing capabilities are opening promising pathways for emotion recognition and regulation, with growing relevance in fields such as affective computing, neurorehabilitation, and human-computer interaction. Through four exploratory studies, this thesis investigates how virtual reality, biofeed back, and machine learning can be combined to recognize and regulate users’ emotional states in real time. First, a machine learning pipeline was developed to classify emotional states using physiological signals collected in immersive and non-immersive virtual reality conditions. Results showed that immersion had a limited impact on subjective emotional ratings, while user-dependent models significantly outperformed user-independent ones, highlighting the importance of personalization in emotion recognition. The second study validated this pipeline in individuals with Alzheimer’s, revealing that emotional reactivity is partially preserved across severity levels. Classification models successfully distinguished between emotional states, healthy and Alzheimer’s participants, and even Alzheimer’s severity levels, underscoring the pipeline’s clinical relevance and generalization. The third study introduced a nature-based virtual reality environment, the Virtual Lev ada, which used real-time adaptation to users’ physiological stress levels via a biofeedback mechanism. This study also implemented and evaluated real-time retraining strategies for the stress classification model, addressing temporal drift and improving model robust ness. Although biofeedback effects were not statistically significant, both adaptive and non-adaptive groups reported reduced physiological arousal and anxiety, supporting the environment’s calming and restorative potential. Finally, the fourth study improved the adaptive virtual reality system by integrating online stress predictions and online model retraining. Results demonstrated improved prediction stability over time and significant reductions in state anxiety, particularly in individuals with elevated stress levels. In conclusion, these findings validate the feasibility and effectiveness of progressively adaptive, personalized virtual reality systems for emotion recognition and regulation. This work contributes with novel insights into how online physiological monitoring and ML adaptation can enhance emotional self-regulation, offering promising directions for affective technologies development and mental health interventions.
- Massively parallel GPU acceleration of population-based optimization metaheuristics: application to the solution large-scale systems of nonlinear equationsPublication . Silva, Bruno Miguel Pereira da; Lopes, Luiz Carlos Guerreiro; Mendonça, Fábio Rúben SilvaHigh-dimensional problems, such as large-scale Systems of Nonlinear Equations, are challenging due to their complexity and nonlinear solution spaces. Population-based optimization metaheuristics, such as Particle Swarm Optimization and Gray Wolf Optimizer, can offer effective approaches. However, their computational demands often exceed the capacity of traditional methods, particularly when addressing these problems at large scales. To address these challenges, parallelization constitutes a promising strategy. Due to the massive parallel processing capabilities, a Graphics Processing Unit (GPU) is well-adapted to the acceleration of population-based metaheuristic optimization algorithms. Thus, employing GPU parallelism can substantially reduce computational time and enable the solution of larger and more complex problems that would be impractical on conventional Central Processing Units (CPUs). GPU-based parallelization of metaheuristic optimization algorithms faces several challenges due to algorithmic diversity and heterogeneous hardware architectures. Different metaheuristics exhibit distinct computational patterns, memory access requirements, and degrees of inherent parallelism, which complicates efficient mapping to GPU architectures. Moreover, variations in GPU hardware can substantially affect performance, often requiring algorithm-specific adaptations and hardware-aware optimizations to fully exploit GPU resources. This research proposes GPU-based parallelization strategies for population-based metaheuristic algorithms to enhance performance on large-scale, high-dimensional optimization problems. It uses GPU parallelism to manage increasing problem sizes while preserving convergence behavior and solution quality. A central goal is a hardware-agnostic model that enables scalable acceleration across diverse computa tional environments, providing a general framework for GPU-based metaheuristic acceleration applicable to various algorithmic paradigms and problem domains. Experimental results indicate that GPU-accelerated metaheuristics using the proposed framework substantially outperform their sequential counterparts, achieving significant speedups. The framework scaled effectively across ten population-based algorithms and ten benchmark problems of increasing dimensionality, utilizing five GPU models, including both consumer-grade and professional-grade hardware. In multi-GPU tests, the framework exhibited superlinear speedup in certain cases. This study highlights the value of modular, reproducible frameworks for GPU based metaheuristics and provides a base for future research in high-dimensional, computationally intensive optimization.
- User profiling with feature selection and explainability: essays on three case studies across different domainsPublication . Freitas, Diogo Nuno Teixeira; Teixeira Freitas, Diogo Nuno; Dias, Fernando Manuel Rosmaninho Morgado Ferrão; Fermé, Eduardo LeopoldoUser profiling is the process of constructing a structured representation of the user within a system. This representation includes information such as preferences, behaviors, and characteristics. Based on the profile, the system can recommend services and products or, in this work, suggest actions. Machine learning methods are commonly used to this end, as they can identify complex patterns among large numbers of attributes. However, not all attributes are relevant. High-dimensional datasets often contain irrelevant, redundant, or noisy features that obscure valuable patterns and reduce model accuracy. To address this, dimensionality reduction techniques—particularly feature selection—are essential. Equally important is the ability to explain a model’s output, since understanding why a model produces a given outcome builds trust and clarifies which steps can change an undesirable situation. This thesis applies feature selection, explainability, causal discovery, and machine teaching techniques to user profiling. The goal is to support decision-making by identi fying the most relevant features, clarifying causal mechanisms, and ensuring that stake holders understand why recommendations are made. Specifically, we investigate the mRMR (minimum-Redundancy-Maximum-Relevance) method for feature selection, ex amine explainability strategies such as feature importance analysis and counterfactuals, apply causal discovery to map cause-and-effect relationships, and use machine teaching to explore profile simplification. We apply this approach in four domains: (i) Marine litter: developing static profiles to identify those who could benefit from literacy interventions; (ii) Football injuries: building predictive models based on player profile dynamics to forecast risk; (iii) Energy poverty: designing models, using counterfactuals, and applying causal discovery to understand health–poverty links; and (iv) Concept complexity: using machine teaching to study profile simplification. These applications show how profiling can deliver targeted literacy interventions, prevent sports injuries, inform preventive policies in energy poverty, and improve the efficiency of user representations and concept learnability.
- A spatial augmented reality-based biofeedback platform for fitness assessment and intervention in older adultsPublication . Ahmad, Muhammad Asif; Ahmad, Muhammad Asif; Bermúdez i Badia, Sergi; Gouveia, Élvio Rúbio QuintalIn most countries around the world, the population of older adults exceeds that of younger individuals. The ageing population has become a global challenge and is causing an increase in chronic diseases and associated health problems. This age group has been associated with conditions such as reduced mobility, prolonged hospitalisations, elevated morbidity rates, and an increased incidence of falls. A sedentary lifestyle has been correlated with numerous health conditions, making it essential to promote innovative health and fitness programs that encourage physical activity. According to the American College of Sports Medicine (ACSM), individuals should engage in 30 minutes of moderate exercise, five days a week, or 20 minutes of vigorous exercise, three times a week. The rapid growth of Information and Communication Technology (ICT) tools has led to a decrease in physical activity, contributing to various health and fitness issues. Balance assessment tools are used to evaluate postural movements, the risk of falls, and balance impairments. Using virtual reality (VR) technology in conjunction with balanced assessment tools may enhance the ecological validity of the instruments, increase performance and standardisation, and reduce administration time. Moreover, combining immersive virtual reality (IVR) with physical exercise, or exergames, enhances motivation and personalises training, effectively preventing falls by improving strength and balance in older adults. While many exergames and simulation applications have been introduced to help people engage in physical activity, limitations exist in user interest, consistency, achievement, and technology, particularly when targeting older adults and populations with specific deficits. Exergames often lack adaptability to individual fitness profiles and high-fidelity immersive virtual environments. To address these limitations, we have created and validated a biofeedback based spatial augmented reality platform that assesses physical and physiological fitness, as well as balance, in older adults through targeted interventions designed to promote wellness and reduce the risk of falls. Target heart rates are evaluated according to ACSM recommendations. The platform also offers simulations of diverse environments and fitness opportunities tailored to meet the specific needs of various populations. We have implemented standard fitness procedures in a virtual environment to assess lower-body strength and cardiorespiratory endurance in older adults.
- Biogeography, population and trophic ecology of cetaceans in a warm-temperate habitatPublication . Ferreira, Rita Borges; Kaufmann, Manfred; Alves, Filipe Marco AndradeCetaceans play a crucial role in marine ecosystems by maintaining their structure and function, providing essential ecosystem services, and acting as sentinel species. Despite the inherent challenges, research efforts to document the distribution and movements of pelagic cetacean populations have increased. Oceanic islands present strategic advantages for studying pelagic cetaceans and integrating datasets from multiple sources is a valuable approach to obtaining the long-term datasets necessary for ecological studies on these long-lived mammals. Ultimately, such integration provides insights into cetacean populations' connectivity and migration patterns across various territories and international borders. The present study aims to elucidate the distribution, movements, and ecological interactions of eight cetacean species, for which limited information exists in the Macaronesia region (Eastern North Atlantic), thereby providing crucial insights for their conservation. This study focuses on understanding the movement patterns and site fidelity of Bryde's whale, sperm whale, and Blainville's beaked whale in the archipelagos of the Madeira, Azores, and Canaries through photographic-identification methods. Additionally, in Madeira, it examines the social structure of the Blainville's beaked whale through photographic identification and analyzes the dietary habits and ecological roles of six odontocete species through stable isotopes. Research findings indicate that Bryde's whales exhibit high site fidelity to the Madeira Archipelago during their seasonal presence, with first-time documented inter archipelago movements between Madeira and the Canaries. This reflects the species' wide habitat range in the Macaronesia region, including international waters. For sperm whales, the study supports the existence of a pelagic population in Macaronesia, with a subset regularly using the Azores and Madeira Archipelagos. Preliminary data from the Canaries suggest a need for further research to evaluate a population significantly impacted by ship strikes. Blainville's beaked whales demonstrate female defense polygyny in the Madeira Archipelago, with strong associations between females and immatures, who exhibit the highest site fidelity rates. These findings underscore the significance of the Macaronesia region in providing essential habitats for these species. The study also highlights variations in dietary preferences and foraging strategies among different species, enhancing our understanding of marine biodiversity and ecosystem dynamics. It emphasizes the need for international collaboration when addressing the conservation challenges of these highly dynamic species. By combining long-term datasets with innovative methodologies, this research significantly contributes to the understanding and protection of cetaceans, underscoring the critical role of oceanic islands in marine conservation efforts.
- Processo educativo: a simulação clínica em enfermagemPublication . Ribeiro, Norberto Maciel; Correia, Fernando Luís de SousaO ambiente para a aprendizagem constitui um dos desafios mais aliciantes na atual geração do século XXI, enriquecida em meios digitais e altamente tecnológicos. Ensinar não representa necessariamente aprender e revela-se um processo complexo, que requer uma estratégia pedagógica promotora de conhecimento. A teoria histórico-cultural de Vigotsky, a fundamentação central do construtivismo crítico de Kincheloe (2006a), a organização concetual do construcionismo de Papert e o uso das ferramentas cognitivas descritas por Jonassen integram referências fundamentais para compreender o fenómeno complexo da aprendizagem e os ambientes promotores de saberes. Salientamos a zona de desenvolvimento proximal da teoria vigotskiana, que constitui uma oportunidade para construir uma aprendizagem mediada pelo professor em cooperação ativa com o estudante. Assim, a simulação clínica em enfermagem destaca-se como estratégia pedagógica e metacognitiva para o desenvolvimento de novas competências, capazes de mobilizar a construção de saberes fundamentais para a melhoria contínua do conhecimento holístico. O estudo desenvolvido na área do Currículo e Inovação Pedagógica, de natureza qualitativa, como método de estudo de caso, numa Escola Superior de Saúde da Região Autónoma da Madeira, procurou compreender o processo educativo da simulação clínica em enfermagem. Na metodologia, utilizaram-se as entrevistas semiestruturadas e o focus group como técnicas de colheita de dados e a análise e interpretação dos dados com a técnica de análise de conteúdo, complementada com a triangulação dos dados colhidos. Compreendemos a simulação clínica em enfermagem como uma metodologia diretiva com recurso a meios tecnológicos e com potencial pedagógico para a construção de uma aprendizagem significativa. A inovação pedagógica no ambiente de simulação é assumida como um designo fundamental em cooperação com a comunidade educativa. Invocamos à ação participativa do estudante, figura central da ação educativa, em todo o processo de simulação clínica, promovendo aprendizagens críticas, criativas, flexíveis, inclusivas e democráticas.
- Very short-range forecasting of wind speed and direction for air traffic operations at Madeira AirportPublication . Alves, Décio Damasceno Mendonça; Alves, Decio; Dias, Fernando Manuel Rosmaninho Morgado Ferrão; Mendonça, Fábio Rúben Silva; Mostafa, Sheikh ShanawazAccurate short-term wind forecasting is essential for aviation safety and efficiency, especially at airports situated in mountainous regions. This doctoral research addresses the challenge of rapidly evolving wind conditions by integrating a newly established network of high-frequency meteorological stations with advanced artificial intelligence models. Through sensor calibration and cost-efficient internet of things infrastructures, the work delivers real-time data at different granular resolutions while reducing measurement error. By employing deep learning approaches combined with feature engineering and data augmentation, the system achieves lower wind speed and direction errors than conventional numerical predictions and supports forecast horizons ranging from one minute to several hours. Interpretable architectures with symbolic methods further enhance trust by revealing how complex terrain influences local wind dynamics and enabling spatial-temporal adjustments of numerical weather predictions. Systematic evaluations confirm improvements in safety-critical metrics for runway operations, including robust classification of wind-induced closures and reduced wind prediction errors. The resulting framework adapts readily to other airports with challenging wind conditions and adheres to international standards for forecasting. In doing so, it strengthens decision-making in air traffic management and fosters greater resilience against disruptive weather events.
- Educação para a proteção civil na Região Autónoma da MadeiraPublication . Freitas, Gregório Magno de Vasconcelos de; de Vasconcelos de Freitas, Gregório Magno; Góis, Liliana Maria Gonçalves Rodrigues de; Pereira, Maria Gorete Gonçalves RochaNeste paradigma de incerteza que vivemos aumenta a sensação de vulnerabilidade e de insegurança nos cidadãos. A segurança, entendida como um direito fundamental, deverá ser prioritária na definição de políticas de proteção civil promotoras da segurança comunitária, atendendo à importância que assume no desenvolvimento social e económico de qualquer país. Urge mais e melhor reflexão sobre este direito, não apenas política, mas, sobretudo, científica, onde a educação democrática deve assumir um papel basilar na construção da cidadania ativa, ambicionando a tão elementar cultura de proteção civil e a resiliência comunitária. No âmbito da sua autonomia, cabe à escola construir um projeto curricular contextualizado. Esta investigação, na área do Currículo e Inovação Pedagógica, assenta num estudo de caso único de natureza qualitativa, uma escola pública dos segundo e terceiro ciclos do ensino básico e ensino secundário da Região Autónoma da Madeira, onde se procura compreender como é que a educação para a proteção civil está a ser implementada. Como técnicas de recolha de dados recorremos à análise documental e às entrevistas semiestruturadas. Como técnicas de análise e interpretação dos dados utilizamos a análise de conteúdo e fizemos, por fim, a triangulação dos dados. Este projeto mereceu a aprovação do Encarregado de Proteção de Dados e da Comissão de Ética da Universidade da Madeira. Verificamos, nos resultados gerais, uma desarticulação entre as dimensões regional, municipal e escolar com o que é internacionalmente preconizado para a elaboração e implementação do currículo para a educação para a proteção civil. Numa comunidade recorrentemente marcada por desastres, exige-se uma prática alicerçada na evidência científica no âmbito da educação e da conceção do currículo para o desenvolvimento comunitário/regional; a implementação dos pilares da escola segura e da educação holística; a redefinição do perfil dos professores; a incrementação dos direitos de toda a comunidade escolar, corresponsável pelos processos de aprendizagens flexíveis, concretizando uma escola justa, inclusiva, democrática, resiliente e aprendente.