Archivio Istituzionale della Ricerca- Università del Salento
Not a member yet
80217 research outputs found
Sort by
Mamba Adaptive Anomaly Transformer with association discrepancy for time series
No full textAnomaly detection in time series poses a critical challenge in industrial monitoring, environmental sensing, and infrastructure reliability, where accurately distinguishing anomalies from complex temporal patterns remains an open problem. While existing methods, such as the Anomaly Transformer leveraging multi-layer association discrepancy between prior and series distributions and Dual Attention Contrastive Representation Learning architecture (DCdetector) employing dual-attention contrastive learning, have advanced the field, critical limitations persist. These include sensitivity to short-term context windows, computational inefficiency, and degraded performance under noisy and non-stationary real-world conditions. To address these challenges, we present MAAT (Mamba Adaptive Anomaly Transformer), an enhanced architecture that refines association discrepancy modeling and reconstruction quality for more robust anomaly detection. Our work introduces two key contributions to the existing Anomaly transformer architecture: Sparse Attention, which computes association discrepancy more efficiently by selectively focusing on the most relevant time steps. This reduces computational redundancy while effectively capturing long-range dependencies critical for discerning subtle anomalies. A Mamba-Selective State Space Model (Mamba-SSM) is also integrated into the reconstruction module. A skip connection bridges the original reconstruction and the Mamba-SSM output, while a Gated Attention mechanism adaptively fuses features from both pathways. This design balances fidelity and contextual enhancement dynamically, improving anomaly localization and overall detection performance. Extensive experiments on benchmark datasets demonstrate that MAAT significantly outperforms prior methods, achieving superior anomaly distinguishability and generalization across diverse time series applications. By addressing the limitations of existing approaches, MAAT sets a new standard for unsupervised time series anomaly detection in real-world scenarios. Code available at https://github.com/ilyesbenaissa/MAAT
Monitoring product circularity and sustainability through Product Lifecycle Management: a bibliometric analysis
Get PDFThe transition to a Circular Economy (CE) in the industrial sector requires the identification and
monitoring of effective Key Performance Indicators (KPIs), in order to measure the effectiveness of circular
strategies in terms of resource efficiency, as well as environmental sustainability. However, today value
chains are characterized by the presence and interactions of several actors, each one involved in different
phases of the Product Lifecycle Management (PLM). Therefore, the identification and availability of
reliable data to measure the selected KPIs can represent a challenge in this process, since data has to be
retrieved from different source of the production chain, requiring effective information sharing approaches.
This need can be addressed through the implementation of effective PLM approaches. To explore the
potentiality of adopting PLM systems to support the monitoring of circularity and sustainability in the
industrial sector, this study presents a bibliometric analysis aiming at highlighting the main trends currently
characterizing this research topic. The results of the analysis show that literature on the topic is moving
from a productivity-centered approach to a sustainability-driven one. However, further research is needed
to provide more qualitative insights on the relevant literatur
Understanding the Psycho-Physiological Impact of Bullying on Adolescents: A Focus on Movement-Based Educational Interventions
No full textEndoscopic papillectomy for ampullary lesions: pooled analysis with meta-regression analysis of outcomes
No full textBackground: Endoscopic papillectomy(EP)is a viable treatment option for ampullary lesions(AL).While many studies have reported low morbidity and acceptable outcomes, early attempts to pool data from these initial experiences have produced conflicting conclusions regarding key technical aspects. To address these uncertainties, we conducted a systematic review and pooled analysis to evaluate the safety and effectiveness of EP for AL,identifying factors that may influence outcomes. Methods: Electronic databases(Medline, Scopus and EMBASE)were searched up to September 2024. Studies that included patients with endoscopically resected AL were eligible. Effectiveness and safety outcomes were pooled by means of a random-effects model to obtain a proportion with a 95% confidence interval(CI). Subgroup analysis, and univariable meta-regression analyses were conducted to explore potential factors affecting outcomes. Results: A total of 61 studies(4,935 lesions)published between 2002 and 2024 were analyzed. Complete resection was achieved in 85.9%of cases, though intraductal involvement limited success. The recurrence rate was 15.2%, however, the majority(92.4%)of patients were managed endoscopically without surgery. The pooled adverse event(AE)rate was 30.0%, with bleeding(12.8%)and post-procedural pancreatitis(11.2%)being the most common complications. Prophylactic stenting reduced pancreatitis risk, while intraductal involvement increased perforation risk.Adjunctive treatments for intraductal involvement posed an increased risk of papillary stricture. Conclusion: Endoscopic papillectomy is a safe and effective treatment for ampullary lesions. However, lesions with intraductal extension pose a higher risk of incomplete resection and perforation, warranting careful evaluation of the benefit-risk balance in these cases.While prophylactic pancreatic stenting may reduce the incidence of post-procedural pancreatitis, optimizing strategies to minimize overall adverse events remains a key focus for future research
xFAIR: A Multi-Layer Approach to Data FAIRness Assessment and Data FAIRification
No full textThe FAIR (Findable, Accessible, Interoperable, Reusable) data principles are crucial for data discoverability, sharing, and
exploitation across diverse contexts, where evaluating data FAIRness reliably is essential for quality assessment and continuous
improvement of data assets. Organizations can identify issues, implement targeted enhancements, and increase data value
and trustworthiness by leveraging actionable insights bluefrom systematic data FAIRness assessment. However, this often
requires heterogeneous metrics and guidelines, particularly when domain-speciic challenges are involved. To bridge the gap
between theoretical FAIR principles and their practical implementation, this paper introduces xFAIR, a multi-layer platform
architecture for assessing and enhancing data FAIRness, and for achieving data FAIRiication in multiple domains. xFAIR
incorporates modules for data acquisition, FAIRness evaluation, and ontology support. Its versatility is demonstrated via three
real-world use cases: i. improving open data portals for Public Administrations, ii. extending FAIR assessment to multi-level
European data portals, and iii. tackling metadata quality in news media by applying FAIR data principles to examine how
source trustworthiness and metadata richness are linked. Each use case highlights speciic aspects (e.g., domain-dependent
metadata validation, or trust scores integration) to enhance quality assurance. Additional components support user feedback
and media literacy. The obtained research outcomes underscore the importance of combining automated metadata validation
with community-driven reinements, to achieve higher data quality and foster ongoing FAIRiication actions across domain
Agents of digitalization: gendered employment patterns and broadband access across Asian economies
No full textWhile much of the literature treats broadband infrastructure as a catalyst for economic growth and in-novation, this study reverses the causal lens by analyzing broadband diffusion as a structural outcome of socioeconomic readiness. Drawing on a unique dataset covering 37 Asian economies from 2002 to 2008, this study develops a mixed analytical framework that integrates macroeconomic drivers and meso-regional structures, specifically the sectoral composition of female employment, as latent indica-tors of digital demand and institutional capacity. This approach merges classical econometric tech-niques – including Robust Least Squares (RLS) and Method of Moments Quantile Regression (MMQR), with advanced Machine Learning (ML) models, such as Generalized Additive Model with Boosting (GAMBoost), GAMBoost for Location, Scale, and Shape (GAMBoostLSS), and Shapley value decomposition. Moreover, Artificial Neural Networks (ANNs) are employed for robustness checks. The empirical findings show that female employment patterns are not merely socioeconomic consequences, but are also crucial in shaping digitalization trajectories. In particular, higher rates of female labor participation in agriculture are negatively associated with broadband penetration. In con-trast, greater female employment in public services and knowledge-intensive sectors predicts stronger adoption of digital infrastructure. These results also interrogate the marginal role of basic education and the paradoxical association between mortality and broadband expansion. By bridging gendered labor patterns, spatial inequality, and digital infrastructure, this study suggests a tailored, ecosystemic perspective on the social structuring of broadband diffusion
The potential of sorghum meal as a replacement of corn meal in the diet for lactating buffaloes: impacts on milk yield and nutrient digestibility
No full textUnder the current climate change scenario and due to its lower irrigation water requirement, sorghum (Sorghum bicolour) meal has been proposed as a viable option in ruminant diet to replace corn meal. The aim of the present study was to evaluate animal performance and in vivo digestibility of lactating buffaloes fed a total mixed ration in which sorghum or corn meal was the primary energy source. Twenty multiparous buffaloes were equally divided into two balanced dietary groups fed the same basal diet supplemented with sorghum meal (Sorghum diet) or corn meal (Corn diet). After a 2–week adaptation period, feed intake and milk yield and components, including somatic cell count and coagulation characteristics, were recorded for 5 weeks. At the end of the experimental period, in vivo nutrient digestibility was assessed using acid insoluble ash as a marker of indigestibility. The use of sorghum meal did not negatively effect on dry matter intake, body weight, or body condition score. Milk yield did not differ between the diets, nor did the content of milk macro components and coagulation characteristics. However, milk urea concentration increased significantly (p < 0.05) in buffaloes fed the sorghum-based diet, indicating reduced dietary nitrogen utilisation. Additionally, these buffaloes exhibited a significant reduction in total tract starch digestibility (−2.8% points, p < 0.05) and a tendency towards lower crude protein digestibility (p = 0.06). We concluded that corn meal can be replaced by sorghum meal in the diet of lactating buffaloes without affecting their productive performance