Anomaly Detection
Learning-based methods for anomaly detection in different applications
Research
We work on the the development of reliable machine learning techniques that can be safely in high risk domains such as healthcare. Although the research questions we try to answer coming from healthcare applications, the methods we develop are generic and transversal making them relevant to application domains sharing the same difficulties. Some current and previous research projects are showcased below:
Quality Control
Performance monitoring in the absence of ground truth
Medical Image Segmentation
Interactive machine learning to circumvent data annotation in healthcare
Health Monitoring
Wearable devides and smart city sensors data for healthcare monitoring
ML Fundamentals
Efficient techniques to optimize the learning process of ML-based systems