Home collection (NIMS)
URI permanente para esta coleção:
Navegar
Entradas recentes
- Reformar as Pensões em PortugalPublication . Grupo de Trabalho para a Reforma da Segurança Social; Bravo, Jorge Miguel; Gomes, Elsa Maria; Gonçalves, João; Castro, Carla; Costa, Vasco; Information Management Research Center (MagIC) - NOVA Information Management School; NOVA Information Management School (NOVA IMS)Portugal encontra-se num dos momentos mais exigentes da história do seu sistema de segurança social, perante um ponto de partida simultaneamente adverso e propício. Adverso, porque as pressões que impendem sobre o sistema se reforçam mutuamente. No plano demográfico, uma das transições mais rápidas da Europa — com fecundidade persistentemente abaixo do limiar de substituição, longevidade crescente e um rácio de dependência de idosos que, segundo o INE, deverá passar de 39 para 73 por cada 100 activos até ao final do século — comprime a base contributiva ao mesmo tempo que pressiona a despesa. No plano económico, um crescimento potencial moderado, uma produtividade inferior à média europeia e a erosão da base contributiva pela informalidade e pela subdeclaração limitam a massa salarial de que o financiamento em repartição depende. No plano social, os baixos salários, as carreiras fragmentadas e uma poupança das famílias reduzida e concentrada em activos imobiliários pouco líquidos deixam grande parte da população idosa dependente quase exclusivamente da pensão pública. Tudo isto num contexto em que, em 2025, a despesa efectiva consolidada dos sistemas públicos de segurança social — Sistema de Segurança Social e Caixa Geral de Aposentações — totalizou 50 267 milhões de euros. Este montante corresponde a aproximadamente 16,4 % do PIB e a 38,4 % da despesa total das Administrações Públicas. Considerando apenas a despesa pública com pensões, o montante ascendeu a 37 589 milhões de euros, equivalente a 12,3 % do PIB e a 28,7 % da despesa pública total. Estas tendências não actuam isoladamente — a insuficiência das prestações contributivas alimenta a procura de apoios assistenciais, e a fraca poupança complementar devolve ao pilar público a responsabilidade que aqueles apoios pretendiam aliviar. Propício, porque o sistema atingiu uma maturidade institucional que permite enfrentar uma reforma fundamentada, ordenada e faseada, e porque subsiste ainda uma janela demográfica — que se fecha em torno de meados da década de 2040 — durante a qual é possível reformar com propósito de forma criteriosa e ponderada, em vez de ter de o fazer mais tarde de forma abrupta e desregulada sob a pressão de uma crise nas finanças públicas. Sobre este pano de fundo, o presente relatório parte de uma constatação e de uma advertência. A constatação é a de que o sistema de pensões desempenha, e continuará a desempenhar, uma função de protecção social e redistributiva essencial, protegendo todos os portugueses, em particular os trabalhadores mais desfavorecidos. A advertência é a de que os indicadores correntes com que essa função tem sido avaliada — em particular os saldos anuais dos regimes contributivos públicos — deixaram, há quase duas décadas, de traduzir fielmente a verdadeira posição financeira e de solvência do sistema, gerando uma ilusão orçamental que adia decisões e transfere encargos, de forma silenciosa mas crescente, para as gerações futuras, e deslegitima o contrato entre gerações.
- Social networks as pathways of political mobility in Portuguese governmentsPublication . Borges, Rafael; Shaul, Carolina; Pinheiro, Flávio L.; Damásio, Bruno; Information Management Research Center (MagIC) - NOVA Information Management School; NOVA Information Management School (NOVA IMS); PLOS - Public Library of SciencePolitical mobility in executive appointments is often explained through party affiliation and career trajectories, yet relational embeddedness across institutional arenas may also characterise those who reach ministerial office. We examine, retrospectively and in an exploratory manner, whether the social networks formed by co-participation in Portuguese governmental and parliamentary venues are associated with ministerial appointments. Using web-scraped data on the composition of the Constitutional Governments and the Portuguese Parliament between 2015 and 2024, complemented with the Registers of Interests of government and parliament officials, we construct an individual-venue bipartite network and its projection onto individuals. We estimate the cumulative network of Governments XXI-XXII and, within each policy-domain, retrospectively screen potential candidates for the XXIII Government by ranking actors in the top five positions of at least one centrality measure (betweenness, reach, closeness, weighted closeness, or eigenvector), computed on portfolio-specific subgraphs and restricted to nodes reachable from the Prime Minister. With the identification of 9 out of 16 in-network ministerial appointments, this procedure outperforms both a random selection baseline (mean recall 0.6%) and a career-ladder heuristic based on prior secretarial experience in the same domain (recall 12.5%), consistent with network centrality being informative about the appointment process in stable, well-populated policy-domains. We assess whether Ministers exhibit systematically higher institutional reach than non-ministerial actors using bipartite degree centrality, visualised via a log-scaled KDE, and validated with non-parametric tests (10,000-iteration permutation tests and Mann-Whitney U tests). We decompose connectivity by venue type (e.g., Government, Parliament, Political Party) using a categorical profile comparison and cosine similarity. Ministers display higher bipartite degree centrality than non-ministerial actors (mean 0.0200 vs. 0.0105; difference 0.0096; permutation p-value = 0.021; Mann-Whitney p-value < 0.001). Categorical profiles diverge substantially (cosine similarity 0.494), with significant gaps concentrated in Government and Parliament venues, while Parliamentary Bodies show no meaningful difference (p-value = 0.919). Venue-specific projected network analysis further reveals that the minister-non-minister centrality gap is largest within the political party network, where ministers score nearly double on eigenvector centrality (+96.0%; permutation p-value = 0.003), pointing to elite positioning within partisan structures.
- Gaussian mixture modeling layer and its application to end-to-end generative high-dimensional clusteringPublication . Marques, Alexandre; Henriques, Roberto; Castelli, Mauro; NOVA Information Management School (NOVA IMS); Information Management Research Center (MagIC) - NOVA Information Management School; Elsevier Science B.V., Amsterdam.Learning multivariate Gaussian mixture models featuring full covariance matrices through backpropagation alone remains a challenge. In this work, we introduce a Gaussian mixture modeling framework that learns general Gaussian mixture distributions over input data through gradient-based backpropagation, enabling local probabilistic modeling within deep neural networks. The forward and backward passes of the proposed implementation exhibit quadratic computational complexity with respect to input dimensionality. The proposed Gaussian mixture modeling layer is differentiable with respect to both the input and its parameters, allowing seamless integration into deep learning architectures. We apply our formulation in an end-to-end generative variational autoencoding approach to high-dimensional image clustering. The empirical results demonstrate the applicability of the Gaussian mixture modeling layer to unsupervised learning through generative modeling, and establish its viability as a building block for deep probabilistic models operating on high-dimensional data.
- Celestial NavigationPublication . Lampreia, Suzana; Policarpo, Hugo; Almeida, Pedro P. de; Roboredo, Nuno R.; Ruivo, Jorge M.; Henriques, Rafael B.; Lobo, Victor; Information Management Research Center (MagIC) - NOVA Information Management School; NOVA Information Management School (NOVA IMS); MDPI - Multidisciplinary Digital Publishing InstituteGlobal Navigation Satellite Systems (GNSS) are incompatible with the underwater domain as radiofrequency signals cannot penetrate the water column, leaving Autonomous Underwater Vehicles (AUVs) reliant on dead-reckoning systems that accumulate positional errors over time. When AUVs surface to reset their navigation, they face another challenge: GNSS itself is increasingly vulnerable to jamming and spoofing in contested environments. Automated Celestial Navigation (CN) has emerged as a promising alternative to other navigation methods, making it possible to derive the absolute position from observations of celestial bodies, entirely independent of human-made signals. This work provides a state-of-the-art review of automated CN technologies, focusing on the literature from 2020 onwards. The review covers Solar Tracking Sensors (STSs), star trackers and horizon detection algorithms and assesses their suitability for AUV integration through a structured SWOT analysis. Following this, a conceptual CN system based on Sunto’s STS is developed for the Light Autonomous Underwater Vehicle platform employing a proposed ten-step integration methodology. Computer-Aided Design models illustrate the conceptual setup, though hydrodynamic/structural verification remains subject to future work. Results suggest that solar-based CN can serve as a periodic absolute position corrector within a hybrid AUV navigation architecture, without requiring satellite infrastructure, which contributes towards a more resilient AUV navigation.
- Exploring Elitism Strategies in Nested Tournament Selection for Multi-Objective Genetic ProgrammingPublication . Pereira, Filipa; Rebuli, Karina Brotto; Giacobini, Mario; Vanneschi, Leonardo; Information Management Research Center (MagIC) - NOVA Information Management School; NOVA Information Management School (NOVA IMS)Nested Tournament (NT) is a Multi-Objective (MO) selection method that enables fine-grained control of selection pressure through sequential single-objective tournaments. Although previously proposed, the impact of elitism strategies within NT remains largely unexplored. This study systematically investigates multiple elitism mechanisms for NT within tree-based MO Genetic Programming (MOGP), including NSGA-II population replacement, crowding distance, first-objective, and a novel ideal-point strategy, comparing them against non-elitist NT and standard NSGA-II. Experiments are conducted with up to five objectives for the accuracy-complexity trade-off and show that elitism design critically influences stability, efficiency, and semantic diversity. Notably, simpler NT-specific elitism strategies achieve comparable performance to NSGA-II at lower computational cost while better preserving semantic diversity. Overall, the findings highlight NT as an efficient MOGP alternative to selection methods based on Pareto Fronts.
- Controlling Functional Complexity for Overfitting Reduction and Improved Interpretability in GPPublication . Silva, Sara; Magessi, Inês Marcão Cortes; Vanneschi, Leonardo; NOVA Information Management School (NOVA IMS); Information Management Research Center (MagIC) - NOVA Information Management School; Institute of Electrical and Electronics Engineers (IEEE)Like other machine learning methods, Genetic Programming (GP) frequently faces the issue of overfitting when applied to supervised learning tasks. Traditional regularization techniques, though well-studied, are challenging to apply to GP due to the free-form nature of the evolved models. This work proposes a novel approach that prevents overfitting while inherently improving the interpretability of GP models. It involves a dual optimization process that minimizes loss while penalizing functional complexity using multi-objective selection mechanisms. The improved complexity measure used in this study approximates the mathematical curvature of a function in linear time. While loss minimization is common in GP, penalizing functional complexity is an additional step aimed at evolving robust and smooth functions, less prone to overfitting and potentially more interpretable. Experimental results demonstrate the effectiveness of the two variants of our method, benchmarked against standard GP and two of the seemingly best overfitting-reduction methods found in the literature. By focusing on both loss and complexity, our approach achieves state-of-the-art generalization on difficult problems and a strong feature selection that improves interpretability, making it a unified improvement of GP.
- Understanding the acceptance and use of artificial intelligence (AI) as a travel planning toolPublication . Dias, Ana Carolina; Tam, Carlos; Oliveira, Tiago; Naranjo-Zolotov, Mijail; NOVA Information Management School (NOVA IMS); Information Management Research Center (MagIC) - NOVA Information Management School; Emarald Group Publishing LtdPurpose This study explores key factors affecting user acceptance and utilization of artificial intelligence (AI) tools, focusing on chatbots used for travel planning. It aims to identify the determinants that shape user behavioral intention and use, offering insights to enhance the design and adoption of AI-driven travel assistance technologies. Design/methodology/approach The extended unified theory of acceptance and use of technology (UTAUT2) model is adapted in combination with AI characteristics. Data were collected from 176 respondents through an online questionnaire to explore the relationships between key variables and behavioral intention, using partial least squares structural equation modeling (PLS-SEM). Findings The results reveal positive correlations between performance expectancy, hedonic motivation, perceived anthropomorphism, perceived intelligence, and the behavioral intention to use AI as a travel planning tool. Besides studying the direct impact, we demonstrate the relevance of examining the indirect impact of features through their mediating role. The study identifies that behavioral intention mediates the relationship between AI-specific characteristics, such as perceived anthropomorphism and perceived intelligence, and the actual use of AI in travel planning. Originality/value This research contributes to the understanding of user behavior in adopting AI for travel planning, highlighting critical factors that can inform AI and tourism stakeholders in designing user-centric tools to enhance travel experiences. We identify habit as a significant determinant of actual AI tool usage, emphasizing its importance for sustained long-term use of AI-powered solutions.
- Designing LLM Agents for Output Checking in Research Data CentersPublication . Ashofteh, Afshin; Carvalho, Ricardo; Campos, Pedro; NOVA Information Management School (NOVA IMS); Information Management Research Center (MagIC) - NOVA Information Management SchoolResearch data centers increasingly face a tension between rapidly growing demand for microdata-driven research and strict confidentiality obligations. Output checking—reviewing tables, models, and descriptive statistics before release from safe centers-remains a key safeguard but is labor-intensive and difficult to scale. This paper presents a governed, semi-automated output-checking architecture that integrates three layers: (i) deterministic rule-of-thumb disclosure checks, (ii) a machine-learning classifier trained on historical release decisions to support triage, and (iii) a large language model (LLM) agent that performs principles-based synthesis and produces structured explanations for auditors and researchers. Using an archival corpus of output requests and released/blocked outputs from a national statistical office, we build a metadata and standardization pipeline for heterogeneous datasets and outputs, engineer disclosure-relevant features, and evaluate a prototype that routes low-risk cases quickly while escalating ambiguous cases for human control. We evaluate the architecture retrospectively on historical output-checking records and distinguish three validation targets: deterministic detection of obvious disclosure risks, ML prediction of institutional outcomes, and LLM-based evidence synthesis for audit support. The contribution is a governance-oriented architecture and empirical prototype for routing and explanation, not an autonomous release mechanism: all unsupported, ambiguous, or high-risk cases remain subject to professional output-checker review.
- High-Frequency Early Warning of Systemic Financial Stress in Europe Using Financial and Non-Financial Data with Machine LearningPublication . Diachkov, Dmytro; Ashofteh, Afshin; NOVA Information Management School (NOVA IMS); Information Management Research Center (MagIC) - NOVA Information Management SchoolEarly detection of systemic financial stress is challenging in fast-moving, nonlinear environments. Traditional early warning systems rely on low-frequency indicators and linear models, limiting their real-time relevance. This paper develops a high-frequency machine-learning framework to monitor systemic financial stress in Europe using daily financial, banking, macro-financial, and sentiment indicators. Diverse data sources are integrated in a strictly time-ordered pipeline to avoid look-ahead bias, with systemic financial stress states defined by threshold exceedances of the ECB's Composite Indicator of Systemic Stress. Feature selection and evaluation are conducted using stability-based regularization, time-series cross-validation, and rare-event metrics. Models combining financial and sentiment indicators outperform linear benchmarks, particularly during periods of intensifying stress. The results suggest that high-frequency machine-learning models can improve the timeliness of systemic stress monitoring and complement existing macroprudential tools.
- A machine learning framework for short-horizon monitoring of systemic financial stress in EuropePublication . Diachkov, Dmytro; Ashofteh, Afshin; NOVA Information Management School (NOVA IMS); Information Management Research Center (MagIC) - NOVA Information Management School; ElsevierEarly detection of systemic financial stress is essential for financial-stability monitoring in environments characterized by rapid market adjustment, nonlinear transmission, and rare stress episodes. This study develops a high-frequency machine-learning framework for short-horizon monitoring of systemic financial stress in Europe using daily financial, banking, interest-rate, exchange-rate, commodity, volatility, and sentiment-related indicators. Systemic financial stress is measured with the European Central Bank’s Composite Indicator of Systemic Stress (CISS), which is transformed into binary stress targets using fixed, rolling-sigma, and rolling-percentile threshold rules. The empirical design evaluates alternative model classes, dataset specifications, feature representations, forecast horizons, and stress definitions within a strictly chronological out-of-sample validation framework that avoids look-ahead bias. Model performance is assessed using rare-event metrics, with primary emphasis on precision–recall AUC, recall, and F1 score. The results show that predictive performance is strongest under the fixed CISS threshold and at short horizons, particularly the 5-business-day forecast window. The preferred specification is a transparent logit model using CISS- and VIX-based predictors with level and change features. Sentiment and uncertainty-related indicators also provide meaningful short-horizon information, while broader standalone market blocks perform more moderately. Nonlinear classifiers are competitive in several settings but do not materially dominate the linear benchmark. Robustness checks show that rolling thresholds preserve the qualitative short-horizon pattern but reduce performance and increase specification instability, especially at longer horizons. Overall, the framework is best interpreted as a disciplined decision-analytics tool for high-frequency systemic-stress monitoring rather than as a threshold-invariant long-horizon stress-prediction model.
- «
- 1 (current)
- 2
- 3
- »