I am interested in developing and applying computational intelligence solutions to real world problems, with a focus on bioinformatics and biomedical engineering, using algorithms that are deeply inspired by how our brain works.
Over the years, great part of my work comprised machine learning applied to neural data and brain-machine interfaces. I have also worked with computational modelling of neural structures that are related to Parkinson's Disease. I'm now particularly interested on machine learning/AI methods applied to bioinformatics and biomedical engineering problems.
Topics of interest include: computational intelligence, bioinformatics, machine learning, signal processing
Ongoing graduate student projects
Intelligent decision support system for breast cancer classification based on mammography images (PhD candidates: Raphael Torres and Nickerson Ferreira. Funding: CAPES)
Breast cancer is the leading cause of cancer-related deaths of women in Brazil. In this project, we are first building a novel mammography dataset working closely with one of the leading cancer institutes in Brazil, the "Liga Norte Riograndense Contra o Câncer". Then, using tools from machine learning, such as deep convolutional networks and federated learning, we aim at designing a system that can eventually support breast cancer diagnosis.
Machine learning for Parkinson's Disease prediction based on gut microbiome (PhD candidate: Marcela de Angelis. Funding: CAPES)
Parkinson's Disease is a neurodegenerative disease that affects more than 3% of people over 65 years old, with figures set to double in the next 15 years. There is still no cure, and current therapies are only able to provide symptomatic relief. Diagnostics is clinical and typically confirmed at later stages of the disease, when motors symptoms become more prominent. In this project, we are investigating machine learning solutions to establish whether the gut microbiome of a person contains relevant information to predict the occurrence of PD.
Ethical and legal constraints on the use of AI in bioinformatics applied to medical decision making (PhD candidate: Luana Ferraz Alvarenga. Funding: CAPES)
The use of AI in bioinformatics is growing, posing risks in sensitive areas like healthcare. Ethical and legal boundaries must be established before widespread clinical use. This project investigates AI in medical decision-making, from the perspective of legislation and ethics, analysing AI approaches, legislation, and real cases to propose guidelines that will contribute towards data protection, bias mitigation, and legislative updates.
Deep Spiking Neural Networks applied to Epileptic Seizure prediction (PhD candidate: Lourena Rocha)
Epilepsy is a chronic neurological condition affecting millions of people worldwide. A major concern is the unpredictability of seizures. This project aims at developing a seizure prediction model using deep spiking neural networks, which have been recently shown to excel in similar tasks while being more energy-efficient. In the future, this may underlie embedded solutions, that could be worn by a person in their daily routine.
Other projects
Lato Sensu Graduate Program in information technology for the health sector (2022 - Present; coordinator)
Funding: Liga Norte Riograndense Contra o Câncer; Unimed Natal; DNA Center
Development of an epidemiological monitoring system for the State of Rio Grande do Norte (2022 - 2023; coordinator)
Funding: Public Health Office from the State of Rio Grande do Norte
Neuro4PD: Neurorobotics Model of Parkinson's Disease (2018 - 2022; Fellow)
Funding: Royal Society The Newton Funding
SINGULARITY - Smart Agent-Based Epidemiological Model Of COVID-19 For Societies With Urban Singularities (2020 - 2022; Co-PI)
Funding: UK Global Challenges Research Fund
SOPHIA - Soft Orthotic Physiotherapy Hand Interactive Aid system (2015 - 2016; Co-PI)
Funding: Royal Society The Newton Funding