Alireza Mousavi-Hosseini

Flow Matching with Semidiscrete Couplings

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

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

On Fitting Flow Models with Large Sinkhorn Couplings

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

Robust Feature Learning for Multi-Index Models in High Dimensions