Automatic generator of scientific abstracts in tourism research
Keywords:Natural language processing, Generation of scientific texts, Tourism, RNN, LSTM, GPT-3
With the new wave of neural networks, architectures based on deep learning have been very successful in the last decade. One of the main applications that these architectures have had is the automatic generation of text, which has recently gained popularity. Different tasks have tried to be solved, such as song generators, chatbots, automatic summaries, and translators, among others. However, few works have focused on generating scientific texts. This paper presents an analysis of the generation of abstracts of scientific articles exploring different types of architectures. 227 articles on Natural Language Processing applied to the tourism sector were collected for this work. With this collection, different types of fine-tuning are proposed where the best result is 0.21 obtained by GPT-3 according to the Jaccard coefficient.
How to Cite
Copyright (c) 2022 Miguel Ángel Álvarez Carmona, Ramón Aranda, Ángel Diaz-Pacheco, José de Jesús Ceballos-Mejía
This work is licensed under a Creative Commons Attribution 4.0 International License.