Generación de un modelo de simplificación automática de textos en español mediante inteligencia artificial

Information and documentation services should be readily available to the public (United Nations,
1994), particularly in vulnerable populations with visual difficulties associated with reading (people
with an intellectual impairment, dyslexia, autism, low scholarity levels, or visual and hearing
impairment, among others). As such, text simplification is a tool for democratizing access to
knowledge by transforming complex writing into a more understandable text. The acquired
knowledge becomes more relevant when it impacts people’s lives, such as when relating to the
spheres of education, politics, or economics.
Nevertheless, text simplification is often time-consuming and demands considerable financial and
human resources, as it requires the participation of professionals with specific knowledge, such as
linguistics or the field pertaining to the texts to be transformed. This project aims to develop a text
simplification model in Spanish in the field of finance utilizing artificial intelligence. A generic dataset
will be constructed to train the Spanish text simplification models; likewise, a dataset of simple and
compound homologous sentences from financial texts to train and assess the automatic
simplification models. Likewise, the non-supervised learning models used for the automatic
simplification, as well as the semi-supervised and supervised models, will be assessed.
The proposed products of this project are the generic dataset utilized to train the Spanish text
simplification models, the simple and compound homologous sentence dataset from financial texts,
and the non-supervised, supervised, and semi-supervised learning models; the results will be
published in two articles in two scientific journals.

Desarrollar un modelo para la simplificación
automática de textos de temáticas financieras en español mediante
inteligencia artificial

Línea de investigación de Ciencias de Lenguaje: Competencias comunicativas en cualquier idioma
y lengua materna
Línea de investigación en Computación: Teoría y Metodologías en Computación