TY  - RPRT
TI  - High-dimensional Asymptotics of Feature Learning: How One Gradient Step Improves the Representation
AU  - Jimmy Ba
AU  - Murat A. Erdogdu
AU  - Taiji Suzuki
AU  - Zhichao Wang
AU  - Denny Wu
AU  - Greg Yang
PY  - 2022
UR  - https://arxiv.org/abs/2205.01445
ID  - 2205.01445
ER  -