Publication: Generación automática de quizzes para museos empleando técnicas de Deep Learning
Loading...
Official URL
Full text at PDC
Publication Date
2021-07-07
Advisors (or tutors)
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Citations
Exportar
Abstract
En este trabajo se presenta el diseño de un algoritmo que permite obtener respuestas incorrectas a preguntas de donde se conoce la contestación correcta, empleando tecnologías de Web Semántica y técnicas de Deep Learning. Se plantea como entorno de explotación un minijuego, que se encuentra desplegado en el Museo de Ciencias Naturales de Madrid. El juego se basa en las áreas de las exposiciones como la evolución de la vida en la Tierra, que abarcan desde los primeros microorganismos hasta el Homo Sapiens, por medio de colección de fósiles, esqueletos, reconstrucciones e ilustraciones que recrean la vida en la Tierra en sus diferentes eras o etapas. Sobre esta base, se plantea la posibilidad de ampliar la programación del juego, en cuanto a un algoritmo que permita definir cuestionarios de tipo pregunta de opción múltiple, en donde habrá una respuesta correcta y tres incorrectas, con posibilidad de establecer niveles de conocimiento. La experiencia irá dirigida a famosos paleontólogos y sitios paleontológicos. Para esto, se emplean tecnologías de la Web Semántica como DBpedia y su variante SpotLigh, Wikidata, Wikipedia, Yago y técnicas de Deep Learning como es el caso de Word2Vect.
This paper presents the design of an algorithm that allows to obtain incorrect answers to questions where the correct answer is known using Semantic Web technologies and Deep Learning techniques. A mini-game, which is deployed in the Museum of Natural Sciences of Madrid, is proposed as an operating environment. The game is based on the areas of exhibitions such as the evolution of life on Earth, ranging from the first microorganisms to Homo Sapiens, through a collection of fossils, skeletons, reconstructions, and illustrations that recreate life on Earth in its different eras or stages. On this basis, the possibility of extending the programming of the game, regarding the methodology for defining questionnaires question multiple choice type, where there will be a correct answer and three incorrect arises. The experience will be aimed at famous paleontologists and paleontological sites. For this, Semantic Web technologies are used as DBpedia and its variant Spotligh, Wikidata, Wikipedia, Yago and Deep Learning techniques as in the case of Word2Vect.
This paper presents the design of an algorithm that allows to obtain incorrect answers to questions where the correct answer is known using Semantic Web technologies and Deep Learning techniques. A mini-game, which is deployed in the Museum of Natural Sciences of Madrid, is proposed as an operating environment. The game is based on the areas of exhibitions such as the evolution of life on Earth, ranging from the first microorganisms to Homo Sapiens, through a collection of fossils, skeletons, reconstructions, and illustrations that recreate life on Earth in its different eras or stages. On this basis, the possibility of extending the programming of the game, regarding the methodology for defining questionnaires question multiple choice type, where there will be a correct answer and three incorrect arises. The experience will be aimed at famous paleontologists and paleontological sites. For this, Semantic Web technologies are used as DBpedia and its variant Spotligh, Wikidata, Wikipedia, Yago and Deep Learning techniques as in the case of Word2Vect.
Description
Trabajo de Fin de Máster en Internet de las Cosas, Facultad de Informática UCM, Departamento de Ingeniería del Software e Inteligencia Artificial, Curso 2020/2021