TY  - RPRT
TI  - Learning to Perturb Hidden Representations for Generalizable Deep Learning
AU  - Hua Li
PY  - 2026
UR  - https://arxiv.org/abs/2605.29525
ID  - 2605.29525
ER  -