An adaptive and gamified educational experience using chatbots
Tutor/a - Director/a
Rodríguez Santiago, Inmaculada
Estudiante
Font Solanich, Laia
Tipo de documento
Projecte Final de Màster Oficial
Fecha
2025
rights
Acceso restringido por decisión del autor
Editorial
Universitat Politècnica de Catalunya
UPCommons
Resumen
This thesis explores the enhancement of an existing gamified educational chatbot system named mIA-UB, through the use of AI techniques such as Natural Language Processing (NLP) and automated planning. The system addresses challenges in educational contexts, including enhancing user engagement through tailored conversational modules and gamification strategies. In addition, the project investigates the integration of Large Language Models (LLMs) to improve chatbot functionality and includes the design of a Hybrid Automated Planning Module (HAPM)using Planning Domain Definition Language (PDDL+) for student activities plans, aiming to improve the chatbot user's productivity. Various modifications were applied to the Natural Language Understanding (NLU) framework of mIA-UB chatbot, enhancing its usability and interactivity. A preliminary usability study with students was also done, analysing interaction efficiency and identifying challenges in user adjustment to dual interaction modes (commands and natural language). We also propose the possible implementation of RAGs as an alternative of the chatbot's NLU framework. Afterwards, the implementation of hybrid adaptive automated planning functionalities took place, enabling task organization in multi-party settings. Two test cases were developed in educational contexts to obtain plans, adapting to the user's needs: one focused on project planning and the other on creating a study plan for an exam. The findings demonstrate the potentiality of adaptive and gamified chatbot systems in educational settings to improve student collaboration, motivation, and task efficiency, while also highlighting areas for further refinement in conversational and planning modules. Additionally, a small prototype showcasing the capabilities of the HAPM has been developed.
Profesorado participante
- Rodríguez Santiago, Inmaculada