Fraunhofer Center for Assistive Information and Communication Solutions – AICOSDescription:
This thesis will focus on dealing with the multimodality of every clinical context. The student will investigate how different types of I can be merged by Machine Learning (and potentially Deep Learning) models to produce more reliable decisions. Also, the dependence between modalities will be studied, important to later focus on potentially missing inputs. Models should be able to produce predictions when some data is missing, although always considering the increasing decision uncertainty. The viability of probabilistic and generative models will be studied for this task.
Outcome:
This thesis aims at the developing of reliable multimodal methods for Machine Learning algorithms in the clinical context. The student will approach the dependency between features and the potential for missing inputs.
Author: Isabel Curioso
Type: MSc thesis
Partner: FCT NOVA – Faculdade de Ciências e Tecnologia da Universidade Nova de Lisboa
Year: 2022