Trabajos de fin de máster (TFM)

URI permanente para esta colección

hdl:123456789/37

Examinar

Examinando Trabajos de fin de máster (TFM) por Palabra clave "1203.17 Informática"

Mostrando 1 - 1 de 1
  • Miniatura
    Publicación
    Meaning aggregation functions
    (Universidad de Educación a Distancia (UNED), 2024-06-23) Lucas Pérez, Gadea; Amigo Cabrera, Enrique; Fresno Fernández, Víctor Diego
    In the contemporary landscape of artificial intelligence and computational linguistics, Natural Language Processing (NLP) systems play a crucial role in understanding, interpreting, and generating human language. This work addresses the significant challenge posed by the black-box nature of deep learning models and the complexity of natural language, particularly the issue of polysemy. To address this challenge, one promising avenue is the concept of semantic distributional representation, which maps texts into a multidimensional semantic space. This approach enhances the visibility and manipulability of linguistic representations. In this work we introduce two novel semantic functions, fspec(v1, v2) and fgen(v1, v2), designed to specialise and generalise the concepts encapsulated by word vectors, respectively. Our research involves defining these functions, characterising their properties, developing an evaluation benchmark, and conducting a comprehensive comparison of candidate functions. The results indicate that while the sum function is most effective for specialisation, polysemy remains a significant source of noise in both specialisation and generalisation tasks. We propose future research directions, including the exploration of multilingual datasets and more sophisticated models to handle polysemy. The advancements from this research hold practical implications for improving the accuracy and applicability of NLP systems in various domains.