Entradas recentes
Development of Self-Cleaning Cementitious Panels with Nano-TiO₂ and Micro-ZnO: Aesthetic and Photocatalytic Impacts of Epoxy Resin Application
( 2025 ) Lessa, Fabíula Pereira; Júnior, Orlando Sousa Lima; Margalho, Élida Melo; Zadeh, Behzad Zahabi; Lima, Vítor Manuel Cunha; Cunha, Eduardo Nunes; Camões, Aires; Rocha Segundo, Iran Gomes da; Carneiro, Joaquim A. O.
Self-cleaning coatings based on nanomaterials like TiO₂ and ZnO are promising for building facades and heritage surfaces. However, their long-term effectiveness is challenged by particle loss due to rain, wind, and abrasion. Epoxy resin can potentially improve particle adhesion but may affect photocatalytic performance and visual appearance. This research aims to develop self-cleaning cement-based panels functionalized with TiO₂ and TiO₂/ZnO, and to evaluate the impact of epoxy resin on their photocatalytic activity and aesthetic properties. The photocatalytic coatings contributed to self-cleaning, promoting a significant increase in photocatalytic efficiency without causing a significant aesthetic impact on the surfaces. The application techniques, spray and dip coating, showed differences in terms of uniformity and thickness, which influenced the final performance. The use of epoxy resin preserves aesthetics but significantly compromises photocatalytic efficiency—up to 84% in some cases. Further research should explore alternative binders and surface treatments that maintain transparency and catalytic activity. These treatments should also be assessed for their long-term durability and resistance to wear.
Predictive incident detection architectural proposal for local AI in SOC
( 2026 ) Mishra, Shashank; Santos, Henrique
The dissolution of the traditional network perimeter and the adoption of zero-trust architectures have inundated
modern Security Operations Centers (SOCs) with high-velocity, heterogeneous telemetry, creating highly stochastic
environments where Advanced Persistent Threats (APTs) thrive by mimicking benign behavior over extended
temporal horizons. Traditional machine learning Intrusion Detection Systems (IDS) rely on monolithic, static
models that suffer from severe performance degradation due to continuous concept and feature drift, label
scarcity, and structural graph bottlenecks. To resolve these operational gaps, this thesis proposes AEGIS-CT, a
highly scalable, self-supervised, and predictive IDS pipeline that mathematically bifurcates the continuous physics
of network traffic from the discrete semantic reasoning of attacker intent. The architecture is structured across
three specialized phases: Phase 1 standardizes incomplete, mixed-type streaming data into a unified latent space
using a Gaussian Copula module and autonomously mitigates structural and semantic drift via an unsupervised
detector (ADWIN-U) coupled with Representation Enhancement for Concept Drift Adaptation (ReCDA). Phase 2
contextualizes anomalies by utilizing an Ontology-Grounded Retrieval-Augmented Generation (OG-RAG) framework,
mapping telemetry to a strict STIX and AttacKG+ cybersecurity knowledge graph to extract unambiguous, semantic
macro-states without relying on numerical data evaluation by Large Language Models (LLMs). Phase 3 integrates
these bifurcated payloads into an Episodic Continuous-Time Temporal Graph Neural Network (CT-TGNN). By
leveraging Time-Aware Batch-Normalized Ordinary Differential Equations (TA-BN-ODE) and decoupling continuous
computation from historical memory via a Global State Registry, the system evaluates temporal network evolution
without adjacency matrix memory explosion. Rigorous theoretical evaluation establishes strict mathematical
bounds on computational complexity, demonstrating the reduction of point process scaling from O(n2) to
O(n3/2). A logical dry-run against the SolarWinds supply-chain compromise validates the pipeline’s theoretical
capacity to prevent adversarial drift poisoning, bridge massive temporal dormancy gaps, and translate complex
topological deviations into highly explainable, sample-level alerts designed for seamless Human-AI teaming.
Ambient intelligent and machine learning applications for increasing efficiency and skills: a case study on secondary education on Math
( 2026 ) Bezerra, Rita Maria Araújo; Novais, Paulo; Durães, Dalila
O desempenho académico no ensino secundário é influenciado por fatores académicos, comportamentais e contextuais. Neste contexto, a integração da Inteligência Artificial, da Aprendizagem Automática e dos Grandes Modelos de Linguagem na educação oferece oportunidades para identificar padrões de desempenho, apoiar a aprendizagem personalizada e facilitar a tomada de decisões pedagógicas baseadas em evidências. Esta tese propõe uma Arquitetura de Referência para a Educação Inteligente, enquadrada numa perspetiva de Inteligência Ambiental centrada no software e orientada por dados. A Inteligência Ambiental é entendida como um paradigma que integra dados educativos heterogéneos, informação contextual e componentes computacionais inteligentes para apoiar a tomada de decisões educativas. A arquitetura desenvolvida integra modelação preditiva baseada na fusão precoce de dados, seleção de atributos e Inteligência Artificial Explicável, bem como um Tutor de Matemática baseado em Geração Aumentada por Recuperação e num Modelo de Linguagem de Grande Dimensão executado localmente através do Ollama, alinhado com o currículo do 10.º ano. A validação centrou-se no desempenho preditivo, na explicabilidade e nas perceções de utilidade e usabilidade, envolvendo, no caso do tutor, um professor e 27 alunos do 10.º ano. Os resultados constituem uma validação inicial de prova de conceito e não incluem a medição longitudinal de ganhos formais de aprendizagem. De um modo geral, a investigação integra modelação preditiva, explicabilidade e tutoria adaptativa numa arquitetura modular, contribuindo para o desenvolvimento de ambientes educativos inteligentes baseados em dados e para a aplicação responsável da Inteligência Artificial em contextos de ensino formal.