معرفی
Caterina Urban is a Research Scientist at INRIA, affiliated with École Normale Supérieure in Paris, where she is a member of the ANTIQUE research team. She holds a PhD in Computer Science from École Normale Supérieure and completed her undergraduate and graduate studies at Università degli Studi di Udine with summa cum laude honors. She previously conducted postdoctoral research at ETH Zurich.
- PhD in Computer Science, École Normale Supérieure, Paris, France (2015)
- MSc in Computer Science, Università degli Studi di Udine, Italy (2011)
- BSc in Computer Science, Università degli Studi di Udine, Italy (2009)
Her research spans formal methods, abstract interpretation, and static analysis, with a strong emphasis on ensuring reliability, fairness, and robustness in machine learning and data science software. She leads projects such as Lyra (on data science software), Libra (fairness certification), ForML, and SAIF. Her work bridges theoretical foundations with practical tool development, targeting real-world software challenges in AI safety.
Her recent publications reveal a consistent focus on applying abstract interpretation to machine learning, data leakage, fairness, and program termination. Key themes include certified robustness, input usage analysis, and perfectly parallel verification, demonstrating both theoretical depth and practical impact. She has developed open-source tools like Libra and FuncTion.
- Best Poster Award at SOFSEM 2011
- Artifact Evaluation Badges at OOPSLA 2020
- Mentored award-winning undergraduate research (Global Undergraduate Awards 2022)
She advises multiple PhD and postdoctoral researchers and has supervised numerous master’s and bachelor’s students. She is actively involved in the academic community, serving on program committees for PLDI, OOPSLA, and CAV, and is the general chair of iFM 2025. She teaches advanced courses at MPRI and international summer schools on abstract interpretation and formal methods.
She leads research on data science software analysis, fairness in ML, and static verification of neural networks, often in collaboration with institutions across Europe. Her lab, part of the ANTIQUE team, fosters interdisciplinary work combining formal methods with AI and software engineering.
