Transformer: o poder da atenção no auxílio ao diagnóstico de múltiplas doenças neurodegenerativas

Abstract

Neurodegenerative diseases (NDDs) cause, among other symptoms, gait instability, have an incurable nature,and present a long and challenging diagnostic process. For this reason, several studies have investigated gait using artificialintelligence models as an alternative to assist in the diagnosis of these diseases. This study presents the main resultsobtained during an undergraduate research project using an innovative method for detecting NDDs, using an Encoder-OnlyTransformer combined with gait analysis in a multi-class classification task. The results achieve high accuracy values andindicate a promising alternative for NDD identification.