TY  - RPRT
TI  - Logarithmic High-Probability Regret for Online Convex Optimization with Two-Point Bandit Feedback
AU  - Haishan Ye
PY  - 2026
UR  - https://arxiv.org/abs/2603.25029
ID  - 2603.25029
ER  -