Alireza Naderi

Alireza Naderi is a Research Fellow in the Department of Mathematics at the University of Oxford. His research interests include the theory of deep learning and random matrix theory. Naderi has contributed to several publications, including works on the spectral analysis of rank collapse in transformers and the convergence of singular values in Gaussian matrices. He has also collaborated on research involving sparse and low-rank neural networks and compressed sensing with generative priors. Naderi has served as a tutor and teaching assistant for courses on optimisation for data science and theories of deep learning.

Alireza Naderi is a PhD student at the University of Oxford, specializing in Deep Learning Theory, High-Dimensional Probability, and Random Matrix Theory. His research explores the theoretical foundations of deep learning, focusing on phenomena such as rank collapse, signal propagation, and the behavior of neural networks under high-dimensional settings. He investigates connections between compressed sensing, generative priors, and Gaussian processes, with applications in signal recovery and machine learning. His work also includes contributions to random matrix theory, such as the convergence of singular values in products of Gaussian matrices. Naderi's research bridges probabilistic methods with machine learning, offering insights into the structure and performance of deep models.