Master Defense of Marcos Felipe de Menezes Mota

28 February 2020
10:00 Master's Defense IC 2 auditorium 3
Theme
Harena System for Medical Training: Causal Modeling for Procedural Content Generation of Clinical Cases
Student
Marcos Felipe de Menezes Mota
Advisor / Teacher
Andrà © Santanchè
Brief summary
Preparing medical students to provide emergency care is one of the biggest challenges in health education, as novice doctors must develop the broad spectrum of knowledge of a general practitioner in the shortest possible time. Researches in medical training demonstrate that a computer system for problem-based online learning, presenting clinical cases to be solved by students, has several benefits in the learning process. Despite the benefits of these educational systems, also known as e-Learning systems, one factor that impedes the widespread adoption of this type of platform is the difficulty and time required to create content suitable for learning. A possible solution would be the generation of this content in an algorithmic way, this is the objective of an area called procedural content generation (GPC). Most of the publications and available medical data report cause and effect relationships between symptoms, diseases and medications, but GPC techniques do not incorporate any formalism of causal inference into their models. Therefore, the approach of this research is to use GPC and causal inference to generate narratives of clinical cases. To materialize this approach, the Harena system was used, a system developed at the Information Systems Laboratory (LIS) in partnership with the Faculty of Medical Sciences (FCM). Harena is a web system that allows the creation of medical narratives in a systematic way and the compilation of narratives in an interactive clinical case. Using the system and models of cause and effect, it was possible to generate playable clinical narratives and with variable narrative schemes. Therefore, although incipient, the results indicate the ability to generate clinical narratives for medical training systems through GPC and causal modeling. Thus, this research recommends further exploration of GPC in the context of medical applications and that the integration with causality models can bring a richer model for GPC algorithms.
Examination Board
Headlines:
André Santanchè IC / UNICAMP
Dario Cecilio Fernandes FCM / UNICAMP
Hélio Pedrini IC / UNICAMP
Substitutes:
Julio Cesar dos Reis IC / UNICAMP
Thiago Martins Santos FCM / UNICAMP