Exploring gender bias in Spanish deep learning models

  1. Ismael Garrido-Muñoz 1
  2. Arturo Montejo-Ráez 1
  3. Fernando Martínez-Santiago 1
  1. 1 Universidad de Jaén, Campus Las Lagunillas
Libro:
SEPLN-PD 2022: Annual Conference of the Spanish Association for Natural Language Processing 2022: Projects and Demonstrations
  1. Miguel A. Alonso (ed. lit.)
  2. Margarita Alonso-Ramos (ed. lit.)
  3. Carlos Gómez-Rodríguez (ed. lit.)
  4. David Vilares (ed. lit.)
  5. Jesús Vilares (ed. lit.)

Editorial: CEUR Workshop Proceedings

Año de publicación: 2022

Páginas: 44-47

Tipo: Capítulo de Libro

DIALNET GOOGLE SCHOLAR lock_openAcceso abierto editor

Resumen

This paper presents a data visualization tool developed during the investigation of the bias present in deeplearning language models in Spanish. The tool allows us to explore in detail the outcome of the response ofthe models we present with a set of template sentences, allowing us to compare the behavior of the modelswhen the templates are presented with a context that alludes to a man or a woman. The exploration of thedata in the tool is performed at various levels of detail, from visualizing the model output itself with itsweights to visualizing the aggregation of the results by categories. It will be this last visualization that willprovide some interesting conclusions about how the models perceive mainly women by their bodies and menby their behavior

Fuente de los datos: Dialnet