ADDI
Not a member yet
47763 research outputs found
Sort by
Cognate facilitation effect on verb-based semantic prediction in L2 is modulated by L2 proficiency
No full textPublished online on 12 December 2024We tested whether verb-based prediction in late bilinguals is facilitated when the verb is a cognate versus non-cognate. Spanish–English bilinguals and Chinese–English bilinguals (control) listened to English sentences such as “The girl will adopt the dog” while viewing a scene containing either a dog and unadoptable objects (predictable condition) or a dog and other adoptable animals (unpredictable condition). The verb was either a cognate or non-cognate between SpanishandEnglishandneveracognatebetweenChineseandEnglish.Bothgroupsofbilinguals were more likely to look at the target (the dog) in the predictable versus unpredictable condition. However, only low-proficient L1 Spanish bilinguals showed greater and earlier prediction when the verb was cognate than when it was non-cognate, suggesting that cognate facilitation effect occurs not only on the cognate word itself but al soon prediction based on this cognate word, and that this effect is modulated by L2 proficiency.This study was funded by the Start-Up Grant from the National University of Singapore (#A-8000008-00-00), the Basque Government through the BERC 2022-2025 program and by the Spanish State Research Agency through BCBL Severo Ochoa excellence accreditation CEX2020001010-S. CDM received funding from the European Research Council (ERC) undertheEuropeanUnion’sHorizon2020researchandinnovationprogramme (Grant Agreement No: 819093) and the Spanish Ministry of Economy and Competitiveness (PID2020-113926GB-I00). AB received the support of a fellowshipfromthe“LaCaixa”Foundation(ID100010434).Thefellowshipcodeis LCF/BQ/DR23/12000006.TheauthorsthankCarlySummerlotforhelpwiththe stimuli recording, Jiawen Ma and Xinxian Zhao forhelpwith thedatacollection and Daiwen Gong for help with preparing the Appendix
MARTYKÁNOVÁ, Darina, Los ingenieros en España. El nacimiento de una élite, Servicio Editorial de la Universidad del País Vasco, Bilbao, 2023, 341 pp.
No full textCiertas consideraciones sobre la relación entre el arbitraje comercial internacional y la tutela jurisdiccional en el ámbito de la Unión Europea
No full textEste TFG realiza una comparativa entre la tutela procurada por los tribunales y el arbitraje comercial internacional, acotado al ámbito de la Unión Europea y a través del análisis de sentencias del Tribunal de Justicia de la Unión Europea. Para ello, se analizarán diferentes cuestiones controvertidas en la práctica entre ambas disciplinas
Ionic materials for thermal energy storage: design, synthesis, characterization and application scope
No full text304 p.In the context of the current energy crisis, thermal energy storage (TES) is gaining renewed importance.Efficient heat storage and release not only facilitate the integration of renewable sources such as solarand wind into the grid but also improve industrial efficiency and reduce dependence on fossil fuels.Among TES technologies, latent heat storage (LHS) using phase change materials (PCMs) stands out forits ability to store and release large amounts of energy within narrow, nearly constant temperatureranges.This doctoral thesis, ¿Ionic Materials for Thermal Energy Storage: Design, Synthesis, Characterizationand Application Scope¿, explores the design, synthesis, and evaluation of organic and hybrid organic¿inorganic ionic materials as PCMs. Owing to their ionic nature, these compounds offer high tunabilitythrough structural modification or ion exchange, allowing their thermophysical properties to be tailoredto specific applications. The materials developed cover an operational range from ~14 °C to 350 °C,enabling use in climate control, electronics thermal management, and industrial waste heat recovery.The work is structured by chemical class and operating temperature:1) Organic ionic plastic crystals (OIPCs) with solid¿solid transitions (>100 J/g) for applications below 45°C.2) Organic¿inorganic hybrids with enthalpies up to 167.5 J/g and stability near 90 °C.3) Dicationic salts (>100 J/g) compatible with conductive additives, operating around 125 °C.4) Alkanoate-based ionic liquids with melting enthalpies up to 207.8 J/g and decomposition above 380°C, competitive with sodium nitrate yet less corrosive and safer around 300 °C.A comprehensive structural and thermophysical characterization was performed¿using NMR, FTIR,XRD, TGA, DSC, hot disk thermal conductivity, density, and corrosion tests¿along with cyclingexperiments (up to 500 cycles) to evaluate durability.Overall, this thesis establishes a foundation for developing new ionic PCMs across a wide temperaturespectrum, offering efficient, tunable, and safer alternatives for thermal energy storage and contributingto a more sustainable and secure energy transition
La implementación de la inteligencia artificial en la auditoría
No full text[spa] La finalidad del presente trabajo consiste en analizar los riesgos y beneficios asociados a la implementación de la inteligencia artificial en el ámbito de la auditoría. Para ello, se ha llevado a cabo una revisión bibliográfica basada en fuentes académicas y profesionales, empleando un enfoque cualitativo. El trabajo se estructura en tres partes. Por un lado, en la primera parte se presenta algunas de las principales ramas de la IA, identificando cuáles de ellas se emplean en el sector de la auditoría y se presenta ejemplos de usos de la inteligencia artificial en los procesos de auditoría. En la segunda parte, se exponen los beneficios y riesgos y medidas de mitigación de los riesgos más significantes. Y, por último, se muestran dos ejemplos de aplicaciones hechas por grandes firmas auditoras que contienen IA. El trabajo evidencia que la implementación de la IA genera muchos beneficios y riesgos operativos, pero en realidad esta herramienta en el sector de la auditoría implica una transformación que va más allá, como un cambio en el mercado laboral. Entre las limitaciones se encuentra como principal limitación la falta de información disponible sobre la IA en la auditoría debido a que se trata de una implementación muy reciente en este sector. El valor del trabajo radica en ofrecer una visión actualizada sobre qué tan integrado está la inteligencia artificial en la auditoría.[eng] The purpose of this paper is to analyse the risks and benefits associated with the implementation of artificial intelligence in the field of auditing. To this end, a literature review has been carried out based on academic and professional sources, using a qualitative approach. The paper is structured in three parts. The first part presents some of the main branches of AI, identifying which of them are used in the auditing sector and presenting examples of uses of artificial intelligence in auditing processes. The second part presents the benefits and risks and mitigation measures for the most significant risks. Finally, two examples of applications made by large audit firms containing AI are presented. The paper shows that the implementation of AI generates many operational benefits and risks, but in reality, this tool in the audit sector implies a transformation that goes beyond this, such as a change in the labour market. Among the limitations, the main limitation is the lack of available information on AI in auditing due to the fact that it is a very recent implementation in this sector. The value of the paper lies in providing an updated view on how integrated artificial intelligence is in auditing
Explorando la frontera: Aplicaciones generativas de la IA en el análisis del comportamiento de los consumidores en línea
No full textJEL: M31[EN] This paper presents a systematic review of the application of generative artificial intelligence (AI) in online consumer behavior analytics (OCBA). With the advent of e-commerce and social media, consumer behavior increasingly occurs online, generating vast amounts of data. This shift necessitates advanced analytical tools, and generative AI emerges as a pivotal technology. Generative AI, distinct from traditional AI, can autonomously generate new content based on learned data patterns, offering innovative approaches to OCBA. Based on the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) methodology and data synthesis method proposed by Webster and Watson (2002), this study analyzes 28 peer-reviewed papers, focusing on how generative AI is applied in OCBA and how it can enhance OCBA performance. The findings show that generative adversarial networks (GANs) are the most used, followed by variational autoencoders (VAEs) and autoregressive models. This review categorizes the application areas of generative AI in OCBA and examines how these technologies enhance OCBA’s effectiveness and efficiency. Furthermore, the paper discusses the challenges associated with generative AI, emphasizing the need to consider ethical issues, such as bias and data privacy. This comprehensive review contributes to a deeper understanding of generative AI’s role in OCBA, outlining its applications and functionalities from a technical perspective. It guides future research and practice, highlighting areas for further exploration and improvement in leveraging generative AI for consumer behavior analytics.[ES] Este artículo presenta una revisión sistemática de la aplicación de la IA generativa en el Análisis del Comportamiento del Consumidor Online (OCBA). Con la llegada del comercio electrónico y las redes sociales, el comportamiento de los consumidores se produce cada vez más en línea, lo que genera enormes cantidades de datos. Este cambio requiere herramientas analíticas avanzadas, y la IA generativa emerge como una tecnología fundamental. La IA generativa, distinta de la IA tradicional, puede generar de forma autónoma nuevos contenidos basados en patrones de datos aprendidos, ofreciendo enfoques innovadores a la OCBA. Basado en la metodología PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) y el método de síntesis de datos propuesto por Webster y Watson (2002), este estudio analiza 28 artículos revisados por pares, centrándose en cómo se aplica la IA generativa en OCBA y cómo puede mejorar el rendimiento de OCBA. Los resultados muestran que las redes generativas adversariales (GAN) son las más utilizadas, seguidas de los autocodificadores variacionales (VAE) y los modelos autorregresivos. La revisión clasifica las áreas de aplicación de la IA generativa en OCBA y examina cómo estas tecnologías mejoran la eficacia y la eficiencia de OCBA. Además, el artículo analiza los retos asociados a la IA generativa, haciendo hincapié en la necesidad de tener en cuenta cuestiones éticas como la parcialidad y la privacidad de los datos. Esta revisión contribuye a una comprensión más profunda del papel de la IA generativa en la OCBA, esbozando sus aplicaciones y funcionalidades desde una perspectiva técnica
Citizenship in the context of contested nationalism: Insights from Basque social movements
No full textThis paper contributes to knowledge about social movements' visions of citizenship. The aim is twofold: on the one hand, to offer an analysis of how social movements understand the subjective and objective dimensions of citizenship. On the other, to explore how a context of national conflict shapes activists' perspectives on this concept. To this end, it presents a case study focused on the Basque Country, a territory in which sub-state nationalism is a central feature of the social and political landscape. The paper is based on a qualitative analysis of the discourses of a number of Basque political activists engaged in campaigning around the issues of feminism, environmentalism, anti-racism and migrant rights, and the movement for the revitalisation of the Basque language. The main conclusion is that the national question adds an additional layer of complexity to the perspectives on citizenship articulated by social movements, especially in relation to language and national identities.Ministerio de Asuntos Económicos y Transformación Digital, Gobierno de España MINECOG17/P1
Ipar Euskal Herriko hots dardarkariak: Hazparneko libertimendua aztergai
No full textArtikulu honetan, Ipar Euskal Herrian gertatzen ari den dardarkarien aldakortasunari erreparatu diogu. Dardarkari ubularra dardarkari hobikaria ordezkatzen ari da eta ikusiko dugu ordezkatzea honezkero guztiz gauzatua den edo ez. Aldakortasuna ez dela berria erakusten duten lekukotasunak aipatzearekin bat, eskura ditugun erreferentzia eta baliabideekin egungo egoera islatu nahi izan dugu.
2024ko Hazparneko (Lapurdi) libertimenduko (16-30 urte arteko) zirtzilen jardunek osatutako corpusean Hazparneko hiztunek dardarkariak ubular edo hobikari ahoskatzen dituzten ikertu dugu. Emaitzek erakutsi dute Hazparneko gazteen artean dardarkari ubularra orokorra izanik ere, libertimendu emanaldiko zenbait egoera eta pertsonaia antzezteko hobikaria erabiltzen dutela. Beraz, ezin uka daiteke Hazparneko hiztun gazteek dardarkari hobikaria hautemateaz gain, nahita ahoska dezaketela. Nahiz eta egunerokoan eurek ez erabili, berariaz ebaki dezaketen hotsa izaten jarraitzen du, komunitatean neurri eta modu batean presente dagoen hotsaren oihartzun edo ispilu gisa
Parsimonious machine learning for the global mapping of aboveground biomass potential
No full textAdvances in computational power and methods, and the widespread availability of remote sensing data have driven the development of machine learning models for estimating global carbon storage. Current models often rely on dozens of predictor variables to estimate aboveground biomass density (AGBD), resulting in accurate but complex models that are challenging to interpret from a biological and ecological standpoint. Yet, it remains unclear whether such model complexity is essential to achieving accurate predictions. This manuscript investigates the potential to create a simpler, yet accurate, global AGBD model. Our approach leverages only climate-based predictors, using a systematic predictor selection process to determine the optimal subset of variables that maximize model accuracy. Surprisingly, we found that a minimal model trained with only four bioclimatic variables outperformed more complex models. When compared to a state-of-the-art complex model and ground-based data, our model achieved comparable accuracy using only four predictors, far fewer than the 186 predictors used in the complex model. In conclusion, we present a lightweight, interpretable climate-based model for AGBD estimation, with the additional advantage of being adaptable for projecting AGBD under future climate scenarios.DB's work is funded by the IKUR programme (https://www.science.eus/en/ikur). This research is supported by María de Maeztu Excellence Unit 2023-2027 Ref. CEX2021-001201-M, funded by MCIN/AEI /10.13039/501100011033; and by the Basque Government through the BERC 2022-2025 programme. Computational analysis was carried on in the infrastructure of the DIPC Supercomputer Center (https://dipc.ehu.eus/en/supercomputing-center
Predicción y análisis de incendios forestales
No full textEste Trabajo de Fin de Grado presenta el desarrollo y evaluación de modelos de aprendizaje profundo para la predicción del riesgo de incendios forestales utilizando imágenes aéreas y satelitales del conjunto de datos FireRisk. Se han entrenado y comparado distintas arquitecturas, tanto redes convolucionales clásicas como modelos basados en atención global-local. Se han analizado métricas clave como la accuracy, la precision, el recall y el F1-score para seleccionar los modelos con mejor rendimiento. Además, se han propuesto mejoras específicas para las arquitecturas ConvNeXt y MaxViT. Los resultados obtenidos proporcionan una base sólida para futuras investigaciones en la automatización y optimización de la predicción del riesgo de incendios en entornos naturales.Gradu Amaierako Lan honetan, FireRisk datu-multzoarekin lan eginez, baso-suteen arriskua
aurreikusteko sakoneko ikasketa bidezko modeloak garatu eta ebaluatu dira. Arkitektura
desberdinak entrenatu eta konparatu dira, hala nola sare konboluzional klasikoak eta arreta
global-lokala uztartzen duten modeloak. Accuracy, precision, recall eta F1-score bezalako
metrikak erabili dira errendimendu oneneko modeloak hautatzeko. Horrez gain, ConvNeXt
eta MaxViT arkitekturetako hobekuntza espezifikoak proposatu dira, datu-multzoaren ezaugarrietara egokitzapena eta generalizazioa lehenetsiz. Lortutako emaitzek oinarri sendoa
eskaintzen dute baso-suteen arriskua modu automatikoan eta eraginkorrean aurreikusteko
etorkizuneko ikerketetarako.This final degree project presents the development and evaluation of deep learning models for
predicting forest fire risk using RGB images from the FireRisk dataset. Various architectures
have been trained and compared, including traditional convolutional networks and models
with global-local attention mechanisms. Key metrics such as accuracy, precision, recall, and
F1-score have been analyzed to identify the best-performing models. Specific improvements for
the ConvNeXt and MaxViT architectures have also been proposed, prioritizing generalization
and data-specific adaptation. The results provide a solid foundation for future research aimed
at automating and optimizing wildfire risk prediction in natural environments