Murat A Erdogdu

From Information to Generative Exponent: Learning Rate Induces Phase Transitions in SGD

When Do Transformers Outperform Feedforward and Recurrent Networks? A Statistical Perspective

Distributional Training Data Attribution: What do Influence Functions Sample?

Learning Multi-Index Models with Neural Networks via Mean-Field Langevin Dynamics

Robust Feature Learning for Multi-Index Models in High Dimensions

A Geometric Analysis of PCA

Learning quadratic neural networks in high dimensions: SGD dynamics and scaling laws

Categorical Distributional Reinforcement Learning with Kullback-Leibler Divergence: Convergence and Asymptotics

A Separation in Heavy-Tailed Sampling: Gaussian vs. Stable Oracles for Proximal Samplers

On the Efficiency of ERM in Feature Learning

Beyond Labeling Oracles: What does it mean to steal ML models?