Hello!
I am a 3rd-year PhD student in the ELLIS VANDAL lab at Politecnico di Torino, Italy. I am supervised by Professor Barbara Caputo and advised by Dr. Marco Ciccone.
My research focuses on Federated Learning (FL) and the development of algorithms to handle data distributions shifts across clients, with a specific interest in computer vision tasks.
I received my Master's and Bachelor's degree in Computer Engineering from Politecnico di Torino with maximum grade, completing them in 2020 and 2018 respectively. It was during my Master's thesis titled Towards Real World Federated Learning that I gained my initial hands-on experience with FL. Under the guidance of Prof. Barbara Caputo (Politecnico di Torino), Prof. Fabio Galasso (Sapienza University of Rome) and Dr. Massimiliano Mancini (University of Tübingen), we modeled the interactions among groups of similar clients through a graph-based approach.
I am a proud member and past President of the Mu Nu Chapter of IEEE Eta Kappa Nu, the honor society of IEEE.
Soon to be Visiting Student Researcher at Stanford University under the guidance of Prof. Sanmi Koyejo!
Research interesets
Artificial Intelligence (AI) has the potential to revolutionize numerous domains, but concerns regarding fairness and responsibility have emerged as critical considerations. Developing AI algorithms that are fair, accountable, and socially responsible is imperative for mitigating biases and ensuring equitable outcomes. In my current research, I focus on leveraging Federated Learning (FL) as a powerful paradigm to reach such goals. FL aims to learn a global model from disparate users' data, while respecting the privacy regulations in force. Looking at realistic scenarios, my main efforts aim to
- Learn models providing fair predictions to the entire data distribution, e.g., not biased towards some less represented demographic groups.
- Broaden the horizons of applicability of FL to real-life vision tasks, such as autonomous driving, or image geo-localization, while introducing realistic constraints, e.g., the absence of labeled data.
For more info and published papers, please visit the Research page.