Nuri Mert Vural

pp.jpg

I’m a final-year PhD student in Computer Science at the University of Toronto, advised by Murat A. Erdogdu, and affiliated with the Vector Institute. Previously, I completed an M.S. in Electrical and Electronics Engineering at Bilkent University, advised by Serdar Kozat.

My research lies at the intersection of optimization and statistics. My PhD work focuses on the first-order optimization methods in high-dimensional settings, with an emphasis on neural network training. I have also worked on stochastic optimization, online learning, bandit algorithms, and non-convex optimization.

Email: vural (at) cs (dot) toronto (dot) edu.

I will be joining the Simons Institute at UC Berkeley as a postdoc with Peter Bartlett and Bin Yu in Fall 2026!

Selected works
  1. Gérard Ben Arous, Murat A. Erdogdu, Nuri Mert Vural, Denny Wu. Learning Quadratic Neural Networks in High Dimensions: SGD Dynamics and Scaling Laws. NeurIPS 2025.
  2. Nuri Mert Vural, Murat A. Erdogdu. Pruning is Optimal for Learning Sparse Features in High Dimensions. COLT 2024.