TY  - RPRT
TI  - Gradient-Flow Optimization as Dynamic Random-Effects Inference: Testing and Early Stopping with Applications to Deep Learning
AU  - Minhao Yao
AU  - Ruoyu Wang
AU  - Xihong Lin
AU  - Lin Liu
AU  - Zhonghua Liu
PY  - 2026
UR  - https://arxiv.org/abs/2605.27991
ID  - 2605.27991
ER  -