Kumar, Akshay

3 publications

TMLR 2025 Early Directional Convergence in Deep Homogeneous Neural Networks for Small Initializations Akshay Kumar, Jarvis Haupt

JMLR 2025 Towards Understanding Gradient Flow Dynamics of Homogeneous Neural Networks Beyond the Origin Akshay Kumar, Jarvis Haupt

TMLR 2024 Directional Convergence near Small Initializations and Saddles in Two-Homogeneous Neural Networks Akshay Kumar, Jarvis Haupt