TY  - RPRT
TI  - Gradient-Free Noise Optimization for Reward Alignment in Generative Models
AU  - Jeongsol Kim
AU  - Hongeun Kim
AU  - Jian Wang
AU  - Jong Chul Ye
PY  - 2026
UR  - https://arxiv.org/abs/2605.11347
ID  - 2605.11347
ER  -