Persona: Sánchez de Castro Fernández, Alejandro
Cargando...
Dirección de correo electrónico
asanchez@lsi.uned.es
ORCID
Fecha de nacimiento
Proyectos de investigación
Unidades organizativas
Puesto de trabajo
Apellidos
Sánchez de Castro Fernández
Nombre de pila
Alejandro
Nombre
3 resultados
Resultados de la búsqueda
Mostrando 1 - 3 de 3
Publicación A Novel Methodology for Enhancing Cross-language and Domain Adaptability in Temporal Expression Normalization(Massachusetts Institute of Technology Press, 2025-04-19) Sánchez de Castro Fernández, Alejandro; Araujo Serna, M. Lourdes; Martínez Romo, Juan; Agencia Estatal de Investigación (España); Universidad Nacional de Educación a Distancia (UNED)Accurate temporal expression normalization, the process of assigning a numerical value to a temporal expression, is essential for tasks such as timeline creation and temporal reasoning. While rule-based normalization systems are limited in adaptability across different domains and languages, deep-learning solutions in this area have not been extensively explored. An additional challenge is the scarcity of manually annotated corpora with temporal annotations. To address the adaptability limitations of current systems, we propose a highly adaptable methodology that can be applied to multiple domains and languages. This can be achieved by leveraging a multilingual Pre-trained Language Model (PTLM) with a fill-mask architecture, using a Value Intermediate Representation (VIR) where the temporal expression value format is adjusted to the fill-mask representation. Our approach involves a two-phase training process. Initially, the model is trained with a novel masking policy on a large English biomedical corpus that is automatically annotated with normalized temporal expressions, along with a complementary hand-crafted temporal expressions corpus. This addresses the lack of manually annotated data and helps to achieve sufficient capacity for adaptation to diverse domains or languages. In the second phase, we show how the model can be tailored to different domains and languages using various techniques, showcasing the versatility of the proposed methodology. This approach significantly outperforms existing systems.Publicación RoBERTime: un nuevo modelo para la detección de expresiones temporales en español(Sociedad Española para el Procesamiento del Lenguaje Natural, 2023-03) Sánchez de Castro Fernández, Alejandro; Araujo Serna, M. Lourdes; Martínez Romo, JuanTemporal expressions are all those words that refer to temporality. Their detection or extraction is a complex task, since it depends on the domain of the text, the language and the way they are written. Their study in Spanish and more specifically in the clinical domain is scarce, mainly due to the lack of annotated corpora. In this paper we propose the use of large language models to address the task, comparing the performance of five models of different characteristics. After a process of experimentation and fine tuning, a new model called RoBERTime is created for the detection of temporal expressions in Spanish, especially focused in the clinical domain. This model is publicly available. RoBERTime achieves state-of-the-art results in the E3C and Timebank corpora, being the first public model for the detection of temporal expressions in Spanish specialized in the clinical domain.Publicación Cross-Domain and Multilingual Temporal Information Extraction(Universidad Nacional de Educación a Distancia (España). Escuela Internacional de Doctorado. Programa de Doctorado en Sistemas Inteligentes, 2026-02-23) Sánchez de Castro Fernández, Alejandro; Araujo Serna, M. Lourdes; Martínez Romo, Juan