Repositório Institucional da Universidade Estadual da Paraíba
Not a member yet
17921 research outputs found
Sort by
On the Carian inscription from Mengefe
No full textThis article provides a new analysis of the recently published Carian inscription from Mengefe. It critically discusses the interpretation of the edition, demonstrates that it is mistaken at crucial points (in identifying the words for ‘grandfather’, ‘grandson’, and ‘and’ as well as a relative clause), and offers a more fitting translation (including an [appurtenance] adjective and a clitic correlative conjunction or disjunction), addressing neglected aspects of the inscription as well (a possible postposition and the word dividers)
BMI as a predictor of progression from TB infection to active TB in PLHIV
No full textBACKGROUND: Low body mass index (BMI) is a globally important risk factor for TB progression. Little is known about this association in people living with HIV (PLHIV) and the functional form of the BMI-TB incidence curve. METHODS: Secondary analysis of a randomised controlled trial of TB preventive therapy among PLHIV in South Africa, Mozambique, and Ethiopia. Participants received 3 months of weekly high-dose rifapentine-isoniazid given once or twice over a period of 2 years. Multivariable fractional polynomials (MFPs) were used to investigate functional forms of BMI. Time to incident TB was modelled using Cox's proportional hazard regression. RESULTS: A total of 76 TB events were documented, giving an overall TB incidence rate of 1.2 per 100 person-years (95%CI 1.0-1.6). Baseline BMI = 30 kg/m(2) was associated with a lower hazard of TB (aHR 0.5, 95% CI 0.2-1.0). Continuous and categorical BMI showed weak evidence of quadratic dose-response relationships (P = 0.08 and P = 0.09, respectively). MFP analysis was consistent with a decline in TB incidence for increasing BMI to around 25 kg/m(2), followed by a less steep decline in TB incidence for increasing BMI >25 kg/m(2). CONCLUSIONS: In PLHIV, BMI showed an inverse log-linear association with TB incidence. The MFP approach showed that the relationship is more complex than a simple log-linear association
Identifying coordination and continuity of care indicators in population-based cancer screening: A Delphi study
No full textBackground: Continuity of care and care coordination are critical issues in virtually all healthcare systems. European guidelines for the quality of screening programs for breast and colorectal cancer describe process, structure, and outcome indicators, but none specifically evaluate coordination and continuity of care during the cancer screening process. Objective: The aim of this study was to identify indicators reflecting care coordination and continuity in population-based breast and colorectal cancer screening program. Methods: A two-round Delphi study was conducted in a sample of 30 cancer screening nurses to identify quality indicators. Results: The final Delphi consensus resulted in six core indicators for the cancer screening program: adequacy of the referral of the target population from the screening program to other health services, waiting time for referral to other health services, understanding of the screening program by professionals involved in the process, effective information flow between professionals involved in the process, participants' perception of the screening program, and, finally, participants' understanding of information about the program. Conclusion: The evaluation of indicators is crucial for quality improvement and should allow a measuring system to be established that would allow a comparison of outcomes for all population-based cancer screening program
La formación profesional dual como mecanismo de creación de empleo juvenil
No full textEl desempleo juvenil en el mundo es uno de los principales problemas. En Europa, y especialmente en España, las tasas son muy altas y existe, desde hace mucho tiempo, una situación de total desajuste entre el sistema educativo y el mercado de trabajo. Para poder solventar esta cuestión, es necesario apostar, según el cuarto objetivo de desarrollo sostenible, por el aprendizaje permanente, que en nuestro país tiene como principal herramienta la formación profesional dual. El análisis de la formación profesional dual, su evolución y situación actual, la comparación con otros sistemas europeos, el contrato para la formación y aprendizaje como mecanismo de inserción en el mercado de trabajo, los aspectos formativos y las cuestiones relativas a la Seguridad Social forman parte de este análisis que concluye con unas propuestas
The amendment of the European Stability Mechanism
No full textThe depth of the economic crisis that began in 208 led to financial stabilization instruments in the European Union (EU). After a period of transience and provisionally, the European Stabilization Mechanism (ESM) allowed the Economic and Monetary Union (EMU) countries to consolidate a permanent financial assistance fund. In 2017, the Commission proposed revising the ESM to revoke its international organizational character and transform it into an EU agency included in the EU's institutional structure. The strong opposition of some Member States (MS) has avoided this possibility. The ESM Treaty has been revised and signed by the Eurozone states and maintains its intergovernmental nature. In any case, some new functions have been added, including its establishment as the Single Resolution Fund (SRF) backstop facility only in case of extreme need and when its liquidity is insufficient. Creating this support to the SRF is essential for consolidating the Banking Union
1. Aprenem a pronunciar
No full textAquest vídeo forma part d’un conjunt de cinc, que porten per títol Aprenem a pronunciar. Conté activitats amb frases, un embarbussament i una cançó, que proposen escoltar i aprendre a pronunciar un so consonàntic del català, la doble ela, amb l’ajuda de moviments de la mà, entre altres estratègies.Aquest vídeo està dedicat a l’aprenentatge de la pronúncia d’un so consonàntic del català: la doble ela. Proposa activitats per escoltar-lo i per aprendre a produir-lo. Per focalitzar en l’escolta, la percepció i la producció del so, només hi ha imatge i veu.Projecte lingüístic de la Facultat d'Educació. Universitat de Barcelon
On the relativistic impulse approximation for the calculation of Compton scattering cross sections and photon interaction coefficients used in kV dosimetry
No full textWe calculate differential and integrated cross sections for the Compton interaction as well as massattenuation (μC/ρ), mass energy-transfer (μCtr/ρ), and mass energy-absorption (μen/ρ)coefficients, within the relativistic impulse approximation (RIA) using Compton profiles (CPs)obtained from unrestricted Hartree–Fock electron densities. We investigate the impact of usingmolecular as opposed to atomic CPs on dosimetric photon interaction coefficients for air, waterand graphite, and compare our cross sections to the simpler Waller–Hartree (WH) andKlein–Nishina (KN) formalisms. We find that differences in μC/ρ and μCtr/ρ resulting from thechoice of CPs within the RIA are small relative to the differences between the RIA, WH, and KNcalculations. Surprisingly, although the WH binding corrections seem accurate when consideringμC/ρ, there are significant discrepancies between the WH and RIA results when we look at μCtr/ρ.The WH theory can differ substantially from the predictions of KN and the RIA in the tens of keVrange (e.g. 6%–10% at 20 keV), when Compton scattering becomes the dominant interactionmechanism. For lower energies, the disagreement further grows to about one order of magnitudeat 1 keV. However, since the photoelectric effect transfers more energy than the Comptoninteraction in the tens of keV range and below, the differences in the total μen/ρ values resultingfrom the choice of Compton models (KN, WH, or RIA) are not larger than 0.4%, and thedifferences between WH and the other two theories are no longer prominent.</p
Dos modelos de especialización ganadera en el último siglo: el Vallès Oriental, 1920-2020
No full textA lo largo del siglo xx, el crecimiento urbano y los cambios en las pautas alimentarias impulsaron una expansión de la producción ganadera, especialmente en las áreas rurales más próximas a las grandes ciudades donde se concentraba la demanda de alimentos. Ello dio lugar a modelos de especialización ganadera distintos, tanto desde el punto de vista social como ambiental. En este artículo, comparamos dos modelos de producción ganadera que se sucedieron en una comarca vecina de la ciudad de Barcelona antes y después de la Revolución Verde, atendiendo a sus consecuencias económicas, sociales y ambientales: uno con explotaciones tradicionales aun básicamente orgánicas y otro en el que la actividad ganadera se transformó en una producción agroempresarial en granjas de engorde muy desintegradas de su entorno agrícola. Esta evolución muestra la gran transformación del sistema agroalimentario global en el último siglo, y sugiere la necesidad de seguir estudiando ese contraste
Enhancing Computer Vision Models Using Persistent Homology-Based Descriptors
No full textTreballs Finals de Grau d'Enginyeria Informàtica, Facultat de Matemàtiques, Universitat de Barcelona, Any: 2026, Director: Petia Radeva, Bhalaji Nagarajan i Carles Casacuberta[en] Deep Neural Networks (DNNs) often fail to capture the underlying hierarchical topological structures present in images, a limitation particularly critical in domains such as medical imaging, where fine anatomical details are essential. Topological Data Analysis (TDA), and specifically its prominent technique, persistent homology, serves as a robust tool to detect and represent these structures by summarizing them into Persistence Diagrams. This final degree project explores the theoretical framework and foundations required to understand persistent homology and its applications to image analysis. It also explores the applications of persistent homology in machine learning, specifically applied to multiclass image classification tasks using Deep Neural Networks (DNNs). An effective way to integrate topological information into DNNs is to encapsulate it into a feature vector. While many existing methods enhance architectures like Transformers or CNNs by generating topological descriptors based on static information, this work focuses on learnable strategies. Leveraging the recent architecture, PHG-Net (Peng et al. [46]) as a foundation, we employ a framework where homological descriptors are not fixed manually but are adaptively selected and optimized by the network during training. In this way, we achieve greater generalization and flexibility by enabling the model to autonomously learn the optimal persistent features, rather than relying on fixed, manually selected descriptors. Consequently, this work aims to validate and extend this methodology by establishing a comprehensive experimental benchmark on ISIC 2018, a complex medical imaging dataset and CIFAR-100, which includes more general objects. Furthermore, we conduct a qualitative analysis to evaluate the robustness and decision-making confidence of topologically enhanced models. This includes examining the specific regions where the model focuses on using Grad-CAMs, and analyzing the geometric organization of learned features through dimensionality reduction techniques such as PCA, t-SNE, and UMAP.
[ca] Les Xarxes Neuronals Profundes (DNNs) sovint no aconsegueixen capturar les estructures topològiques jeràrquiques subjacents presents a les imatges, una limitació especialment crítica en imatges mèdiques, on els detalls anatòmics fins són essencials. L’Anàlisi Topològica de Dades (TDA), i específicament la seva tècnica més destacada, l’homologia persistent, serveix com una eina robusta per detectar i representar aquestes estructures resumint-les en Diagrames de Persistència. Aquest Treball de Final de Grau explora el marc teòric i els fonaments necessaris per entendre l’homologia persistent i les seves aplicacions a l’anàlisi d’imatges. També s’exploren les aplicacions de l’homologia persistent en l’aprenentatge automàtic, aplicat específicament a tasques de classificació d’imatges utilitzant Xarxes Neuronals Profundes. Una forma efectiva per integrar la informació topologica a les DNNs és encapsular-la en vectors de característiques. Mentre que molts mètodes existents milloren arquitectures com els Transformadors o les CNN, generant descriptors topològics basats en regles estàtiques, aquest treball es centra en estratègies aprenibles. Utilitzant l’arquitectura PHG-Net (Peng et al. [46]) com a base, emprem un marc on els descriptors homològics no es fixen manualment, sinó que se seleccionen de manera adaptativa i són optimitzats per la xarxa durant l’entrenament.
D’aquesta manera, aconseguim una major generalització i flexibilitat permetent que el model aprengui de manera autònoma les característiques persistents òptimes, en lloc de dependre de descriptors fixos seleccionats manualment.
Consegüentment, aquest treball té com a objectiu validar i estendre aquesta metodologia establint un punt de referència experimental exhaustiu amb l’ISIC 2018, un conjunt de dades complex d’imatges mèdiques, i el CIFAR-100, que inclou objectes més generals. A més, realitzem una anàlisi qualitativa per avaluar la robustesa i la confiança en la presa de decisions dels models millorats topològicament. Això inclou examinar les regions específiques on el model focalitza la seva atenció utilitzant Grad-CAMs, i analitzar l’organització geomètrica de les característiques apreses mitjançant tècniques de reducció de dimensionalitat com PCA, t-SNE i UMAP
Unraveling Charge-Transfer States and Their Ultrafast Dynamics in Artificial Light-Harvesting Complexes
No full textPhotosynthesis relies on highly organized pigment–protein complexes in order to store sunlight energy as biochemical energy. These complexes capture light with remarkable efficiency and are responsible for ultrafast charge separation within a finely tuned energy landscape provided by the protein environments, producing one of nature’s most sophisticated energy conversion systems. Inspired by nature, de novo designed proteins have been proven to be versatile platforms to emulate the function of natural light-harvesting complexes and reaction centers. With Stark and ultrafast transient absorption spectroscopies, we explored the exciton and charge-transfer (CT) mixing, as well as the excited-state dynamics, of a chlorophyll a analogue (Zn-pheophorbide a) in dimers formed within 4-α-helix bundles whose design was previously guided by molecular dynamics simulations. Due to dimerization, we observe an increase in the CT character of the excitonically coupled dimers’ excited state in comparison to monomeric ZnP. Furthermore, additional nonradiative relaxation pathways, together with the formation of transient species absent in monomeric systems, were observed for the dimers. We demonstrate that de novo designed proteins can replicate key features of photosynthetic energy conversion, serving as tunable scaffolds for optimizing light-harvesting processes. Ultimately, these systems have promising applications including photovoltaic cells and biomedical treatments based on sustainable materials