43569 research outputs found
Sort by
MetaTrans-FSTSF: a transformer-based meta-learning framework for few-shot time series forecasting in flood prediction
No full textTime series forecasting, particularly within the Internet of Things (IoT) and hydrological domains, plays a critical role in predicting future events based on historical data, which is essential for strategic decision making. Effective flood forecasting is pivotal for optimal water resource management and for mitigating the adverse impacts of flood events. While deep learning methods have demonstrated exceptional performance in time series prediction through advanced feature extraction and pattern recognition, they encounter significant limitations when applied to scenarios with sparse data, especially in flood forecasting. The scarcity of historical data can severely hinder the generalization capabilities of traditional deep learning models, presenting a notable challenge in practical flood prediction applications. To address this issue, we introduce MetaTrans-FSTSF, a pioneering meta-learning framework that redefines few-shot time series forecasting. By innovatively integrating MAML and Transformer architectures, our framework provides a specialized solution tailored for the unique challenges of flood prediction, including data scarcity and complex temporal patterns. This framework goes beyond standard implementations, delivering significant improvements in predictive accuracy and adaptability. Our approach leverages Model-Agnostic Meta-Learning (MAML) to enable rapid adaptation to new forecasting tasks with minimal historical data. Our inner architecture is a Transformer-based meta-predictor capable of capturing intricate temporal dependencies inherent in flood time series data. Our framework was evaluated using diverse datasets, including a real-world hydrological dataset from a small catchment area in Wuyuan, China, and other benchmark time series datasets. These datasets were preprocessed to align with the meta-learning approach, ensuring their suitability for tasks with limited data availability. Through extensive evaluation, we demonstrate that MetaTrans-FSTSF substantially improves predictive accuracy, achieving a reduction of up to 16%, 19%, and 8% in MAE compared to state-of-the-art methods. This study highlights the efficacy of meta-learning techniques in overcoming the limitations posed by data scarcity and enhancing flood forecasting accuracy where historical data are limited
Molecular biology and design research : A new paradigm and transdisciplinary approach to communicating science
No full textThe transdisciplinary ChromDesign project explored and demonstrated the power of Design research for developing new tools for visualizing and communicating complex concepts in molecular biology to diverse audiences.ChromDesign received funding from the European Union’s Horizon 2020 research and innovation program under the Marie Skłodowska-Curie grant agreement No. 813327. The authors are sincerely grateful to Tina Karagyozova, Francesc Ribot, Marc Marti-Renom, Geneviève Almouzni, Natàlia Dave, and Luciano Di Croce for their thoughtful feedback, which has helped enrich the manuscript. Our thanks also go to all partners in ChromDesign—PhD fellows, supervisors, trainers, communication and research managers—whose collaborative efforts have made the network not only innovative but also fun and engaging
Compositionality for hierarchical reinforcement learning
No full textRecent breakthroughs in AI have proven that reinforcement learning can be used successfully to solve complex sequential decision problems to achieve near-optimal solutions. However, despite the successful applications that use function approximation techniques, RL algorithms face open challenges that involve data efficiency and generalization. In this line, hierarchical methods have been historically applied to simplify learning as much as possible. In this thesis we study how algorithms can exploit compositionality properties to obtain the optimal solution along with a hierarchical decomposition of a large problem. To do this, we devise algorithms in which the agent works at different levels of abstraction to learn a series of base behaviors. These can be combined to optimally solve a higher level task specification. Such base behavior can be reused in different ways to bring computational efficiency by reducing the number of learning samples used to solve each problem.Los avances recientes en IA han demostrado que el aprendizaje por refuerzo puede utilizarse con éxito para resolver problemas complejos de decisión secuencial y lograr soluciones casi óptimas. Sin embargo, a pesar de las aplicaciones exitosas que emplean técnicas de aproximación de funciones, los algoritmos de aprendizaje por refuerzo enfrentan desafíos abiertos que incluyen la eficiencia y la generalización. Para eso, los métodos jerárquicos se han aplicado históricamente para simplificar el aprendizaje lo más posible. En esta tesis, estudiamos cómo los algoritmos pueden explotar las propiedades de composicionalidad para obtener la solución óptima junto con una descomposición jerárquica de un problema grande. Para ello, diseñamos algoritmos en los que el agente trabaja en diferentes niveles de abstracción para aprender una serie de comportamientos base. Estos pueden combinarse para resolver de manera óptima una especificación de tarea de nivel superior y pueden reutilizarse de diversas maneras para mejorar la eficiencia computacional.Programa de Doctorat en Tecnologies de la Informació i les Comunicacion
Parallel networks to predict TIMP and protease cell activity of nucleus pulposus cells exposed and not exposed to pro-inflammatory cytokines
No full textBackground. Intervertebral disc (IVD) degeneration is characterized by a disruption of the balance between anabolic and catabolic cellular processes. Within the nucleus pulposus (NP), this involves increased levels of the pro-inflammatory cytokines interleukin 1beta (IL1B) and tumor necrosis factor (TNF) and an upregulation of the protease families matrix metalloproteinase (MMP) and a disintegrin and metalloproteinase with thrombospondin motifs (ADAMTS). Primary inhibitors of these proteases are the tissue inhibitors of matrix metalloproteinases (TIMP). This work aims at contributing to a better understanding of the dynamics among proteases, TIMP, and pro-inflammatory cytokines within the complex, multifactorial environment of the NP. Methods. The Parallel Network (PN)-Methodology was used to estimate relative mRNA expressions of TIMP1–3, MMP3, and ADAMTS4 for five simulated human activities: walking, sitting, jogging, hiking with 20 kg extra weight, and exposure to high vibration. Simulations were executed for nutrient conditions in non- and early-degenerated IVD approximations. To estimate the impact of cytokines, the PN-Methodology inferred relative protein levels for IL1B and TNF, reintegrated as secondary stimuli into the network. Results. TIMP1 and TIMP2 expressions were found to be overall lower than TIMP3 expression. In the absence of pro-inflammatory cytokines, MMP3 and/or ADAMTS4 expressions were strongly downregulated in all conditions but vibration and hiking with extra weight. Pro-inflammatory cytokine exposure resulted in an impaired inhibition of MMP3, rather than of ADAMTS4, progressively rising with increasing nutrient deprivation. TNF mRNA was less expressed than IL1B. However, at the protein level, TNF was mainly responsible for the catabolic shift in the simulated pro-inflammatory environment. Overall, results agreed with previous experimental findings. Conclusions. The PN-Methodology successfully allowed the exploration of the relative dynamics of TIMP and protease regulations in different mechanical, nutritional, and inflammatory environments in the NP. It shall stand as a comprehensive tool to integrate in vitro model results in IVD research and approximate NP cell activities in complex multifactorial environments.This work was supported by HORIZON EUROPE European Research Council (ERC-2021-CoG-O-Health-101044828)
Could better-quality employment improve population health? Findings from a scoping review of multi-dimensional employment quality research and a proposed research direction
No full textBackground: Precarious employment, a specific part of the conceptual spectrum of employment quality (EQ), has been established as an important risk to individual and population health and well-being when compared to a standard employment circumstance. There remains a need, however, to explore whether and how EQ might be used as a tool to not only protect but also advance population health and well-being. Methods: The purposes of this scoping review were to assess the analytic treatment of the multiple dimensions of EQ and the stances researchers take to characterize the state of knowledge of EQ that supports the idea that better EQ is a health-promoting factor. Quantitative, qualitative, and mixed-methods primary studies that included at least three of the seven conceptually-informed EQ dimensions were eligible. Studies were assessed for EQ dimensions represented, how dimensions were treated analytically, the pathogenic, ambivalent, or salutogenic stances used by investigators, and what each might tell us about how to leverage aspects of better-quality employment to improve population health. Results: A total of 78 studies were included; 54 of these treated EQ dimensions in an interrelated way. Of the analytically interrelated studies, none had an explicit salutogenic stance. Some evidence suggests that a handful of EQ types might present an equal or reduced risk of poor health than the standard employment relationship, frequently used as a historic gold standard. Conclusion: Research with a salutogenic stance might build our understanding of whether and how employment could be used to advance our collective well-being
El duque de Medinaceli y el partido español (1709-1710)
No full textEl partido español que pretendía forjar una agenda política propia ante Francia experimentó un auge decisivo en los años 1709 y 1710, durante las conversaciones de La Haya y Geertruidenberg. Dos personajes lideraron aquella política: por una parte, el conde de Bergeyck, partidario de negociar con Holanda yGran Bretaña para lograr una paz separada a cambio de compensaciones comerciales en las Indias; por otra, el duque de Medinaceli, ministro de exteriores, hasta que fue detenido y encarcelado en el castillo de Segovia, acusado por Felipe V de conspiración
GENIE Learn: human-centered generative AI-enhanced smart learning environments
No full textThis paper presents the basis of the GENIE Learn project, a coordinated three-year research project funded by the Spanish Research Agency. The main goal of GENIE Learn is to improve Smart Learning Environments (SLEs) for Hybrid Learning (HL) support by integrating Generative Artificial Intelligence (GenAI) tools in a way that is aligned with the preferences and values of human stakeholders. This article focuses on analyzing the problems of this research context, as well as the affordances that GenAI can bring to solve these problems, but considering also the risks and challenges associated with the use of GenAI in education. The paper also details the objectives, methodology, and work plan, and expected contributions of the project in this context.This work was supported by grants PID2023-146692OB-C31, PID2023-146692OB-C32 and PID2023-146692OB-C33 funded by MICIU/AEI/10.13039/501100011033 and by ERDF/EU, project GENIELearn. DHL (Serra Húnter) acknowledges support by ICREA Academia
Pathways to instability: how decreasing oil prices impact political stability in petrostates – Lessons from the 1980s oil glut
No full textExisting research has shown a correlation between lower oil prices and political instability in oil-rich countries. This paper examines the conditions under which declining oil revenues can contribute to political instability in such countries, with a focus on the impact of the 1980s oil glut on 29 oil-rich countries. Employing Qualitative Comparative Analysis (QCA), the study identifies three distinct pathways to political instability during this period. Each instance of political instability corresponded to a distinct causal configuration bringing together low oil abundance and high oil dependence with one specific political condition: either high clientelism, relatively high public service provision, or significant political freedom. The findings illustrate the complex interplay of factors that influence the relationship between declining oil revenues and political stability. While rooted in a historical context, the results offer insights for the future, as the global transition to renewable energy may similarly affect oil revenues, posing potential risks to political stability in oil-rich countries.This work was supported by the Spanish Ministry of Science and Innovation program “Strategic Projects on the Ecological Transition and Digital Transition” and is part of the OILDOWN research project (Grant number: TED2021-132846 A-I00)
