Federico
Di Gennaro.
PhD student · ETH Zürich
I am a PhD student at ETH Zürich, advised by Prof. Fanny Yang in the Statistical Machine Learning Group.
My broad research interests focus on how AI systems can make reliable decisions in open-ended environments, where they need to evolve over time to become better versions of themselves. This broad direction can be divided into the following three main areas:
- Sequential Decision Making
- Agentic Systems
- AI Safety
Building on this premise, my idea is to bridge the best and worst cases of the new generation of AI systems. On the one hand, these models can now aim to solve very complex problems by repeatedly adapting to new feedback and data. On the other hand, increasing autonomy and continuous adaptation without human supervision can give rise to safety failure modes with unpredictable consequences. Explore my research
01 / A little background
Before ETH, I spent spring 2025 at UC Berkeley with Prof. Nikita Zhivotovskiy, and fall 2024 at Università di Milano with Prof. Nicolò Cesa-Bianchi.
I hold a master's in Applied Mathematics from EPFL and a bachelor's in Mathematical Engineering from Politecnico di Milano.
02 / Updates
- New preprint
“Fast Unlearning at Scale via Margin Self-Correction” is now online.
- ICML 2026 · Spotlight
“Hedging on the Frontier: Learning New Tasks with Few Samples” was accepted as a Spotlight paper (top 2.2%)!
- AISTATS 2026
“Efficient Logistic Regression with Mixture of Sigmoids” was accepted to AISTATS.
Earlier updates
- Fall 2025
Joined the Statistical Machine Learning Group at ETH Zürich as a PhD student.
“Instance-Dependent Regret Bounds for Nonstochastic Linear Partial Monitoring” was accepted to NeurIPS 2025!
- Spring 2025
Visiting Student Researcher at UC Berkeley, hosted by Prof. Nikita Zhivotovskiy.
- Fall 2024
Visiting Student Researcher at the University of Milan, hosted by Prof. Nicolò Cesa-Bianchi.