Instituto Superior de Engenharia
URI permanente desta comunidade:
Notícias
A definir... (EST)
http://est.ualg.pt/
Navegar
Percorrer Instituto Superior de Engenharia por Objetivos de Desenvolvimento Sustentável (ODS) "04:Educação de Qualidade"
A mostrar 1 - 10 de 15
Resultados por página
Opções de ordenação
- Automated LLM exploit generation framework (ALEGF)Publication . Guerreiro, Joel; João SantosSoftware development frequently suffer from programming defects, mistakes and design flaws introducing critical security vulnerabilities. These weaknesses can be exploited causing significant operational disruption and organizational losses. Identifying and exploiting such vulnerabilities is a complex task that requires binary or source code detailed analysis, operating systems deep knowledge, file system formats and processor architectures, as well as expertise in multiple techniques. Traditional Automatic Exploit Generation (AEG) approaches rely on symbolic execution, fuzzing and heuristicbased methods. In this paper, an automated framework is presented and designed to evaluate Large Language Models (LLMs) to detect vulnerabilities and generate exploits without human intervention. The framework operates within a controlled, sandboxed and fully reproducible environment to prevent real-word security risks while enabling systematic experimentation. The objective is to assess whether LLMs can serve as auxiliary tools for software security testing and vulnerability analysis. The results provide insights into the feasibility, limitations and potential LLM-driven automated exploitation, yet not competitive with AEG tools state-of-the-art.
- Continual learning for object classification: integrating AutoML for binary classification tasks within a modular dynamic architecturePublication . Turner, Daniel; Cardoso, Pedro; Rodrigues, JoaoFor humans it is quite easy to identify a new object after learning to identify existing ones, but not for a machine. Deep neural networks (DNN) are the foundation of the current state-of-the-art methods for training machines to recognize sets of objects. The issue is that any modification to the DNN weights that were trained to classify an initial set of objects has the potential to seriously impair the network’s ability to make those initial classifications; this behaviour is referred to as catastrophic forgetting (CF). This paper presents a continual learning (CL) architecture that can deal with CF. The architecture is composed of two primary parts: (i) The feature extraction component, which is based on the ResNet50 backbone and (ii) the modular dynamic classification (MDC) component. The latter is made up of multiple sub-networks that gradually assemble themselves into a tree-like structure that reorganizes itself as it learns over time, so that each sub-network can operate independently. The MDC relies heavily on binary classification, and here the application of automated machine learning (AutoML) was introduced, where each binary classifier is tailored on-the-fly, and is/can be different from object to object. The strategy involves a calculated selection from a predefined list of model types and parameters, optimizing them for their respective tasks. Results demonstrate that we advanced the adaptability and performance of the network, emphasizing the transformative potential of AutoML in modular CL approaches. Tests on the CORe50 dataset showed accuracy results of 81.1%, which are above the state of the art for CL architectures.
- CTCovid19: automatic Covid-19 model for computed tomography scans using deep learningPublication . Antunes, Carlos; Rodrigues, Joao; Cunha, AntónioCOVID-19 is an extremely contagious respiratory sickness instigated by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Common symptoms encompass fever, cough, fatigue, and breathing difficulties, often leading to hospitalization and fatalities in severe cases. CTCovid19 is a novel model tailored for COVID-19 detection, specifically honing in on a distinct deep learning structure, ResNet-50 trained with ImageNet serves as the foundational framework for our model. To enhance its capability to capture pertinent features related to COVID-19 patterns in Computed Tomography scans, the network underwent fine-tuning through layer adjustments and the addition of new ones. The model achieved accuracy rates that went from 97.0 % to 99.8 % across three widely recognized and documented datasets dedicated to COVID-19 detection.
- Desenvolvimento rápido de aplicações web em low-codePublication . Noel, Felipe Cardoso Maia; Cardoso, Pedro J. S.Durante meu estágio na empresa Atos tive a oportunidade de trabalhar com tecnologias inovadoras, incluindo a plataforma de desenvolvimento low-code Mendix e a metodologia de e desenvolvimento rápido de aplicações (Rapid Application Development, RAD). O meu trabalho envolveu o desenvolvimento web utilizando a referida plataforma Mendix. Neste contexto, pude adquirir habilidades significativas em criação de aplicativos de negócios eficientes e personalizados, utilizando a abordagem low-code para acelerar o processo de desenvolvimento. Os principais pontos do trabalho desenvolvidos durante o estágio foram: 1. Desenvolvimento de Aplicativos Web: Trabalhei no desenvolvimento de aplicações web personalizadas para clientes de uma empresa multinacional de engenharia (que por motivos de confidencialidade, não possível revelar o nome), utilizando a plataforma Mendix. Esses aplicativos atendiam a diversas necessidades, desde gerenciamento de processos de negócios até soluções de automação. 2. Metodologia RAD: Fui exposto à metodologia RAD, que enfatiza ciclos de desenvolvimento curtos e interativos. Isso me permitiu colaborar de perto com os stakeholders e adaptar rapidamente os aplicativos às suas necessidades em constante evolução. 3. Colaboração Interdisciplinar: Trabalhei em estreita colaboração com uma equipa interdisciplinar, incluindo desenvolvedores, analistas de negócios e designers de UX/UI. Essa experiência me proporcionou uma visão abrangente do ciclo de vida completo de desenvolvimento de software. 4. Customização e Integração: Fui responsável por personalizar os aplicativos Mendix para atender aos requisitos específicos dos clientes. Além disso, integrei esses aplicativos com sistemas legados e outras ferramentas, garantindo uma operação suave e eficiente. Em resumo, o estágio na empresa Atos foi uma oportunidade valiosa para adquirir conhecimento e experiência no desenvolvimento web utilizando a plataforma Mendix e a metodologia RAD. Aprendi a colaborar eficazmente em uma equipa multidisciplinar, a adaptar rapidamente os aplicativos às necessidades dos clientes e a entregar soluções de alta qualidade. Esta experiência fortaleceu minha base de conhecimento em desenvolvimento de software e preparou-me para desafios futuros na área das tecnologias da informação.
- Digital cultural heritagePublication . Portalés, Cristina; Rodrigues, Joao; Rodrigues Gonçalves, Alexandra; Alba, Ester; Sebastián, JorgeMost contemporary thinkers agree that we are going through a time of historical change, building a different concept and model of social interrelation. Our ways of life and work have changed, as have the ways in which we communicate and relate to each other. Likewise, an increasing consensus indicates the need to reconfigure traditional social and cultural structures. The Internet, the virtual social networks, and the Information and Communication Technologies (ICTs) have coalesced into a new collective consciousness—a world intercommunicated from the local to the global [1]. The fusion of tradition, culture, history, and legacy with technology, innovation, and interaction provides an attractive system that serves both as an artistic expression and as a fundamental tool for diffusion in cultural institutions [2]. For instance, the usage of interactive technologies such as virtual reality (VR) or augmented reality (AR), combined with multidimensional or multimodal representations [3], provides a significant novelty. User interaction offers a broader perspective, making people more aware of their actions, helping them become the true center of the application. It also enables interactive artistic expression through alternative realities, as well as narration supported by the use of virtual avatars.
- Do personality traits matter for safety behaviour? The boundary role of safety trainingPublication . Sousa, Cátia; Coelho, AnneSafety behaviour in the workplace is influenced by both individual characteristics and organizational practices; however, the conditions under which these factors interact remain insufficiently understood. Drawing on an interactionist perspective, this study examines whether perceived safety training effectiveness functions as a contextual condition that shapes the influence of personality traits on safety behaviour. A cross-sectional design was adopted, and data were collected through an online questionnaire from 268 workers across diverse professional backgrounds. Measures included safety behaviour, personality traits (neuroticism and conscientiousness), and perceived safety training effectiveness. Data were analysed using descriptive statistics, correlation analyses, multiple regression, and moderation analyses, controlling for age and gender. The results showed that neuroticism was negatively associated with safety behaviour, whereas conscientiousness did not present a significant effect when perceived safety training effectiveness was included in the model. Perceived safety training effectiveness emerged as the strongest predictor of safety behaviour. Importantly, perceived safety training effectiveness moderated the relationship between conscientiousness and safety behaviour, such that its influence was stronger at lower levels of training and diminished as training increased. These findings suggest that perceived safety training effectiveness was associated with a weaker relationship between conscientiousness and safety behaviour. By suggesting that the relationship between personality traits and safety behaviour may depend on organizational conditions, this study contributes to a more nuanced understanding of safety behaviour and highlights the central role of training as a key organizational resource for promoting safer work practices.
- Enhancing MILAGE LEARN+ with Machine Learning to improve students’ performancePublication . Figueiredo, Mauro; rodrigues, jose; Martins, Paula Ventura; Zacarias, Marielba; Milharó, DanielaStudents currently attending school were born after the year 2000 and have grown up surrounded by technology, including smartphones, tablets, the Internet, video games, and social media. Typically, conventional educational activities in schools fail to engage these students, leading many of them to struggle academically. This paper presents the strategy adopted by the free MILAGE LEARN+ platform to address these challenges. Artificial Intelligence supports learning personalization by recommending suitable activities tailored to each student’s individual needs, enabling both lower-performing and higherperforming learners to enhance their academic progress. This paper investigates how Artificial Intelligence, specifically through various machine learning methods, can enrich the learning experience offered by the MILAGE LEARN+ platform. Several machine learning approaches are evaluated and analysed based on data from the platform and student outcomes collected during a Mathematics course.
- From cues to engagement: a comprehensive survey and holistic architecture for computer vision-based audience analysis in live eventsPublication . Lemos, Marco; Cardoso, Pedro; Rodrigues, JoaoThe accurate measurement of audience engagement in real-world live events remains a significant challenge, with the majority of existing research confined to controlled environments like classrooms. This paper presents a comprehensive survey of Computer Vision AI-driven methods for real-time audience engagement monitoring and proposes a novel, holistic architecture to address this gap, with this architecture being the main contribution of the paper. The paper identifies and defines five core constructs essential for a robust analysis: Attention, Emotion and Sentiment, Body Language, Scene Dynamics, and Behaviours. Through a selective review of state-of-the-art techniques for each construct, the necessity of a multimodal approach that surpasses the limitations of isolated indicators is highlighted. The work synthesises a fragmented field into a unified taxonomy and introduces a modular architecture that integrates these constructs with practical, businessoriented metrics such as Commitment, Conversion, and Retention. Finally, by integrating cognitive, affective, and behavioural signals, this work provides a roadmap for developing operational systems that can transform live event experience and management through data-driven, real-time analytics.
- Guest intelligence applications acceptance model: an approach with the UTAUT Model and PLS-SEMPublication . Ramos, Celia; Ashqar, RashedIn a world where innovations are made daily, how people interact with and use technology is becoming increasingly important. It is relevant to investigate how new technologies are used and accepted in this environment. Acceptance assessment has been done in research on creating information systems, considering the UTAUT model and the data analysis technique PLS-SEM. Although this digital environment is pertinent to society in general, it is even more appropriate when interacting with tourists because it may give personalised goods and services. In order to assess acceptance in terms of guest insights proportioned by technologies to recommend personalised services and products by evaluating an application that recommends products and services personalised, considering guests’ intelligence, this article analyses the use and acceptance of a technological application whose characteristics meet the aforementioned.
- Identificação da ocupação e uso do solo com base em imagens provenientes de deteção remota e em algoritmos de machine learning: a Reserva da Faia Brava como caso de estudoPublication . Pacheco, Paula Maria de Fraga Borges; Luís, Joaquim; Loureiro, Nuno de SantosEsta dissertação procura identificar a ocupação e uso do solo com base em imagens de deteção remota e algoritmos de Machine Learning (ML), utilizando como estudo de caso a Reserva da Faia Brava, situada no vale do Côa, distrito da Guarda, Portugal. O estudo avalia a exequibilidade de ferramentas de código aberto, como o QGIS e o plugin Orfeo Toolbox, para implementar fluxos de trabalho de classificação supervisionada. Foram processadas imagens de alta resolução obtidas por UAV (2,7 cm/pixel) e pelo satélite Pléiades-Neo (30 cm/pixel), integrando índices de vegetação (GLI, NDVI, SAVI e MSAVI) e métricas texturais de Haralick. O treino dos modelos foi realizado numa quadrícula com 500 metros de lado, selecionada pela sua diversidade ecológica, e posteriormente testada em outras áreas da reserva para avaliar a capacidade de generalização. Dois algoritmos de ML, Random Forest (RF) e Support Vector Machine (SVM), foram testados, com desempenhos avaliados através de matrizes de confusão, F1-scores e coeficientes Kappa. Os resultados evidenciaram a superioridade dos ortofotomosaicos UAV face às imagens de satélite, especialmente quando combinados com análise textural, embora tenham sido identificadas limitações relacionadas com variações sazonais da vegetação e a interoperabilidade entre sensores. O algoritmo RF mostrou maior consistência enquanto o SVM revelou sensibilidade à complexidade espectral. O estudo destaca a aplicabilidade prática destes métodos para a monitorização ambiental, sublinhando a importância das soluções open source para a democratização das tecnologias de deteção remota. Como produto final foi produzido, com base nos modelos de ML, uma carta temática para a área total da Reserva da Faia Brava, abrangendo quatro classes: árvores e arbustos, vegetação herbácea, afloramentos rochosos e outras ocupações e usos do solo.