After pursuing classical studies at school, Giovanni received a BSc in Computer Engineering from the University of Pavia, an MSc in Machine Learning from Royal Holloway University of London (RHUL), and a PhD in Machine Learning and Security with the EPSRC CDT in Cyber Security at RHUL. Before joining the Turing, he was a postdoctoral fellow at EPFL with an EcoCloud grant.
Giovanni's research interests span the broad area of Machine Learning theory, combined with its application to information leakage estimation for privacy and security. Within this area, he has also worked on new security definitions, and on applications of leakage estimation to side channels and to attacks against Machine Learning models.
He is further interested in Machine Learning techniques with provable guarantees under basic assumptions on the data distribution (e.g., Conformal Predictors).
- Theory, foundations, and privacy-security-fairness-robustness properties of Machine Learning
- Information leakage estimation for security and privacy
- Security/privacy definitions, and alternatives to differential privacy
- Traffic analysis (e.g., website fingerprinting) and defences
- Distribution-free learning for supervised problems and for anomaly detection (e.g., Conformal Prediction)
- Philosophical and mathematical foundations of learning