https://hdl.handle.net/10481/18656 Sun, 05 Apr 2026 12:52:36 GMT 2026-04-05T12:52:36Z https://hdl.handle.net/10481/110904 How to achieve gender parity in science? Providing global evidence on key educational and economic drivers González-Salmón, Elvira; Chinchilla-Rodriguez, Zaida; Robinson García, Nicolás; Nane, Gabriela F. Gender parity in science depends on a complex interplay of social, economic and educational variables. In this study, we compile a longitudinal dataset at the country level combining scientific bibliographic data from Dimensions, with the World Bank Open Data (WBOA), and the UNESCO Institute for Statistics (UIS). Our goal is to identify conditions and pathways that could lead to gender parity in different world regions, by applying time-series forecasting methods (ARIMA and Exponential Smoothing), along with correlation analysis and Bayesian networks. While results vary by region, one recurring recommendation emerging from our models is the need to increase the number of researchers and the percentage of women graduating in Engineering, Manufacturing, and Construction, as this appears to be a critical driver for reducing gender disparities in the scientific workforce. https://hdl.handle.net/10481/110904 https://hdl.handle.net/10481/110902 Inferring Gender from Author Names with Local LLMs: A Multi-Model Evaluation Herrero Solana, Víctor; González-Salmón, Elvira; Robinson García, Nicolás Gender identification of researchers is a common practice in scientometric studies examining inequalities in science. The most widely used approach relies on inferring gender from author names using commercial APIs or name-gender dictionaries, which often lack transparency and reproducibility. This study explores the use of local open-weight Large Language Models (LLMs) as an alternative for name-based gender classification. We evaluate 25 models from seven leading families (Llama, Gemma, Phi, Mistral, Qwen, DeepSeek, and Yi), ranging from 270 million to 70 billion parameters, using a reference dataset of nearly 200,000 names across 195 countries extracted from Wikidata. Results show that top-performing models achieve F1-Scores above 0.93 for both gender categories, positioning local LLMs as a viable, cost-effective, and reproducible alternative to proprietary tools. A critical performance threshold emerges at approximately 7 billion parameters, above which all models achieve acceptable results, with diminishing returns beyond 12-14 billion. All models exhibit systematic gender bias, showing higher precision for men and higher recall for women, indicating a tendency to classify ambiguous names as male. Mistral-Nemo-12b emerges as the optimal choice, balancing accuracy, computational efficiency, and gender equity. https://hdl.handle.net/10481/110902 https://hdl.handle.net/10481/108707 El Servicio de consultoría en investigación: Un enfoque comercial y bibliométrico Torres Salinas, Daniel En este curso, se introduce el servicio de consultoría en investigación, sobre todo de carácter bibliométrico, que toma su base en la Bibliometría Evaluativa, una disciplina consolidada por pioneros como Francis Narin (CHI) y Eugene Garfield (ISI) en los años 70. Se centra sobre todo en el modelo de Servicio/Consultoría (Ej. EC3metrics), donde se prioriza el expertise y el análisis ad-hoc de alto valor estratégico, a diferencia del modelo de Herramienta/Producto. Presentan los servicios que se dirigen a cuatro clientes principales: Instituciones (informes, políticas científicas), Revistas, Investigadores (sexenios y acreditaciones) y Formación. Asimismo, se explica la gestión de proyectos a través de un flujo de cuatro fases: Entrevista Inicial (clave para definir objetivos con gestores), Plan de Trabajo, Ejecución/Seguimiento y Entrega. A continuación, repasan cómo definir el presupuesto a través del Número de horas (perfiles senior y junior), el Coste de los datos (bases de datos y datasets bibliométrico) y el Expertise del equipo. Seguimos con la descripción del perfil del consultor que requiere principalmente competencias Documentación, seguida de Informática y Administración de Empresas, con habilidades emergentes de Data Scientist. Terminamos hablando de la calidad y la confianza que se aseguran mediante la Transparencia Total (procesos y personas), la Certificación ISO 9001 y la estricta Confidencialidad y protección de datos sensibles. https://hdl.handle.net/10481/108707 https://hdl.handle.net/10481/107065 ChatGPT y sus hermanas: Una introducción conceptual a la Inteligencia Artificial Generativa para bibliotecas públicas Torres Salinas, Daniel; Arroyo-Machao, Wenceslao Este curso, impartido en el marco de las XXI Jornadas Bibliotecarias Provinciales de Granada el 3 de octubre de 2025, está orientado a introducir de manera conceptual la inteligencia artificial generativa (IAG) en el entorno de las bibliotecas públicas, con un enfoque práctico y reflexivo para profesionales del sector. Los contenidos integran el funcionamiento de los modelos de lenguaje grandes (LLMs), el manejo de plataformas clave como ChatGPT, Perplexity y NotebookLM, las reglas fundamentales para dominar la IA, la importancia de la alfabetización digital y los desafíos éticos y pedagógicos que plantea la automatización y la búsqueda dinámica de información. El curso combina charlas teóricas y un supertaller práctico, fomentando la exploración de aplicaciones reales, la personalización de herramientas, la colaboración entre sistemas y la adopción de estrategias para aprovechar la IA de forma crítica e innovadora en los servicios bibliotecarios y educativos. https://hdl.handle.net/10481/107065 https://hdl.handle.net/10481/107064 The OpenAlex database in review: Evaluating its applications, capabilities, and limitations Forchino, María Veronica; Torres Salinas, Daniel Abstract: This systematic review examines OpenAlex, focusing on its technical features and its applications within science studies and scientific information sources. The paper offers a critical perspective on OpenAlex’s role in democratizing and promoting equity in access to scientific information. A total of 146 articles published between 2022 and 2025 are analyzed and classified by functionality and application, including AI-based validation processes. Findings reveal the rapid adoption of OpenAlex as an open alternative to commercial databases, particularly across the Global South. However, notable metadata limitations—especially regarding affiliations, languages, and document types—still affect analytical reliability. The study highlights key opportunities and challenges for researchers, libraries, institutions, and the technical community, emphasizing the need for complementary validation and responsible assessment strategies. Maria Veronica Forchino carried out this research with funding from the 2025 International Postdoctoral Mobility Call, between institutions belonging to the Ibero-American University Association for Postgraduate Studies (AUIP), as part of “Action Line 3: Mobility Grants for Postgraduate Studies” of the AUIP 2024-2026 Action Plan. Junta de Andalucía. https://hdl.handle.net/10481/107064