Zhenyu Zhu (2026)

Research Interests

  • Machine Learning
  • Deep Learning Theory

Biography

On August 24th, 2026 I successfully defended my PhD thesis. The thesis, entitled “Theoretical Studies of Learning in Overparameterized Neural Networks” was supervised by Professor Volkan Cevher.

I received my Bachelor degree at University of Science and Technology of China (USTC) in 2019. I then got my Master at EPFL in Computer Science in 2022. My master project was supervised by Prof. Volkan Cevher. In September 2022 I started my PhD at LIONS.

Theoretical Studies of Learning in Overparameterized Neural Networks

Z. Zhu / V. Cevher (Dir.)  

Lausanne, EPFL, 2026. 

Detailed record

View at publisher

Multi-Step Alignment as Markov Games: An Optimistic Online Mirror Descent Approach with Convergence Guarantees

Y. Wu; L. Viano; K. Antonakopoulos; Y. Chen; Z. Zhu et al. 

Transactions on Machine Learning Research. 2025. num. 12/2025.

Detailed record

Training Deep Learning Models with Norm-Constrained LMOs

T. Pethick; W. Xie; K. Antonakopoulos; Z. Zhu; A. Silveti-Falls et al. 

2025. Forty-Second International Conference on Machine Learning, Vancouver, Canada, 2025-07-13 – 2025-07-19.

Detailed record

How Gradient Descent Balances Features: A Dynamical Analysis For Two-Layer Neural Networks

Z. Zhu; F. Liu; V. Cevher 

2025. The Thirteenth International Conference on Learning Representations, Singapore, 2025-04-24-2025-04-28.

Detailed record

Benign Overfitting in Deep Neural Networks under Lazy Training

Z. Zhu; F. Liu; G. Chrysos; F. Locatello; V. Cevher 

2023. 40th International Conference on Machine Learning (ICML 2023), Honolulu, Hawaii, USA, 2023-07-23 – 2023-07-29.

Detailed record

Initialization Matters: Privacy-Utility Analysis of Overparameterized Neural Networks

J. Ye†; Z. Zhu; F. Liu; R. Shokri; V. Cevher 

2023. 37th Conference on Neural Information Processing Systems (NeurIPS 2023), New Orleans, Louisiana, USA, 2023-12-10 – 2023-12-16.

Detailed record

Sample Complexity Bounds for Score-Matching: Causal Discovery and Generative Modeling

Z. Zhu; F. Locatello; V. Cevher 

2023. 37th Conference on Neural Information Processing Systems (NeurIPS 2023), New Orleans, Louisiana, USA, 2023-12-10 – 2023-12-16.

Detailed record

Robustness in deep learning: The good (width), the bad (depth), and the ugly (initialization)

Z. Zhu; F. Liu; G. Chrysos; V. Cevher 

2022. 36th Conference on Neural Information Processing Systems (NeurIPS 2022), New Orleans, USA, 2022-11-28 – 2022-12-03.

Detailed record

Generalization Properties of NAS under Activation and Skip Connection Search

Z. Zhu; F. Liu; G. Chrysos; V. Cevher 

2022. 36th Conference on Neural Information Processing Systems (NeurIPS 2022), New Orleans, USA, 2022-11-28 – 2022-12-03.

Detailed record

Extrapolation and Spectral Bias of Neural Nets with Hadamard Product: a Polynomial Net Study

Y. Wu; Z. Zhu; F. Liu; G. Chrysos; V. Cevher 

2022. 36th Conference on Neural Information Processing Systems (NeurIPS 2022), New Orleans, USA, 2022-11-28 – 2022-12-03.

Detailed record

Controlling the Complexity and Lipschitz Constant improves Polynomial Nets

Z. Zhu; F. Latorre; G. Chrysos; V. Cevher 

2022. 10th International Conference on Learning Representations (ICLR), Virtual, April 25-29, 2022.

Detailed record