I'm a Ph.D. student at the Department of Computer Science & Applied Mathematics, Weizmann Institute of Science, where I'm very fortunate to be advised by Prof. Ohad Shamir. My research is supported by an Azrieli graduate fellowship.
I spent the summer of 2024 at Apple Machine Learning Research led by Samy Bengio, working on privacy preserving ML with Vitaly Feldman and Kunal Talwar.
My main research interests are machine learning theory, optimization, statistical learning and deep learning theory.
Previously, I completed my M.Sc. at Weizmann, for which I received the Dean's prize for outstanding M.Sc. student.
Before that I completed by B.Sc. in mathematics at Tel Aviv University.
In somewhat of a past life, I studied guitar at the Center for Jazz Studies in the Israel Conservatory of Music, was supported by the America-Israel Cultural Foundation and played in several musical groups. I'm still enthusiastic about music.
Beyond Benign Overfitting in Nadaraya-Watson Interpolators
Daniel Barzilai, Guy Kornowski, Ohad Shamir
arxiv preprint. [arxiv]
Differentially Private Bilevel Optimization
Guy Kornowski
Preliminary version at Theory and Practice of Differential Privacy workshop (TPDP) 2025.
arxiv preprint. [arxiv]
Improved Sample Complexity for Private Nonsmooth Nonconvex Optimization
Guy Kornowski, Daogao Liu, Kunal Talwar
International Conference on Machine Learning (ICML) 2025. [arxiv]
Trade-offs in Data Memorization via Strong Data Processing Inequalities
Vitaly Feldman, Guy Kornowski, Xin Lyu
Theory and Practice of Differential Privacy workshop (TPDP) 2025, oral presentation.
Best Paper Award at The Impact of Memorization on Trustworthy Foundation Models Workshop @ ICML 2025.
Conference on Learning Theory (COLT) 2025. [arxiv]
The Oracle Complexity of Simplex-based Matrix Games: Linear Separability and Nash Equilibria
Guy Kornowski, Ohad Shamir
Conference on Learning Theory (COLT) 2025. [arxiv]
First-Order Methods for Linearly Constrained Bilevel Optimization
Guy Kornowski, Swati Padmanabhan, Kai Wang, Zhe Zhang, Suvrit Sra
Neural Information Processing Systems (NeurIPS) 2024. [arxiv]
Efficient Agnostic Learning with Average Smoothness
Steve Hanneke, Aryeh Kontorovich, Guy Kornowski
Algorithmic Learning Theory (ALT) 2024. [arxiv]
An Algorithm with Optimal Dimension-Dependence for Zero-Order Nonsmooth Nonconvex Stochastic Optimization
Guy Kornowski, Ohad Shamir
Preliminary version at NeurIPS at Workshop on Optimization for Machine Learning 2023, oral presentation.
Journal of Machine Learning Research (JMLR) 2024. [arxiv] [JMLR]
From Tempered to Benign Overfitting in ReLU Neural Networks
Guy Kornowski*, Gilad Yehudai*, Ohad Shamir
Neural Information Processing Systems (NeurIPS) 2023, spotlight. [arxiv]
Near-optimal learning with average Hölder smoothness
Steve Hanneke, Aryeh Kontorovich, Guy Kornowski
Neural Information Processing Systems (NeurIPS) 2023. [arxiv]
Deterministic Nonsmooth Nonconvex Optimization
Michael I. Jordan, Guy Kornowski, Tianyi Lin, Ohad Shamir, Manolis Zampetakis
Conference on Learning Theory (COLT) 2023. [arxiv]
Oracle Complexity in Nonsmooth Nonconvex Optimization
Guy Kornowski, Ohad Shamir
Neural Information Processing Systems (NeurIPS) 2021, oral presentation.
Journal of Machine Learning Research (JMLR) 2022. [arxiv] [JMLR]
Finalist for the Best Paper Prize for Young Researchers in Continuous Optimization (2025).
On the Hardness of Meaningful Local Guarantees in Nonsmooth Nonconvex Optimization
Guy Kornowski, Swati Padmanabhan, Ohad Shamir
NeurIPS Workshop on Optimization for Machine Learning 2024. [arxiv]
Open Problem: Anytime Convergence Rate of Gradient Descent
Guy Kornowski, Ohad Shamir
COLT 2024 open problem. [arxiv]
- Update: Solved beautifully by [Zhang, Lee, Du and Chen]!
On the Complexity of Finding Small Subgradients in Nonsmooth Optimization
Guy Kornowski, Ohad Shamir
NeurIPS Workshop on Optimization for Machine Learning 2022, oral presentation. [arxiv]
High Order Oracle Complexity of Smooth and Strongly Convex Optimization
Guy Kornowski, Ohad Shamir
arxiv technical note, 2021. [arxiv]
Since 2021, I volunteer as a machine learning mentor at Magshimim (site in Hebrew). Magshimim is a national education program that provides computer science education to underserved youth, equipping them with knowledge and tools for a future career in tech.