Machine Learning Diferencialmente Privado

With the rapid advancements in AI and machine learning (ML), data privacy concerns have gained renewed attention, reinforced by the rise of global data protection laws. In this landscape, differential privacy (DP) stands out as the gold standard, offering mathematically rigorous guarantees. However, despite the expanding ML literature, many algorithms still lack privacy-preserving mechanisms or dedicated research on data protection. This research project aims to bridge this gap by identifying unexplored areas in (differentially) private ML and proposing initial solutions. Specifically, we focus on Topological Neural Networks (TNNs). Our work aligns with Sustainable Development Goals 3 (Good Health and Well-being) and 10 (Reduced Inequalities) by enhancing privacy in ML systems that increasingly impact society, including areas such as health care.