TY  - RPRT
TI  - Discounted Beta-Bernoulli Reward Estimation for Sample-Efficient Reinforcement Learning with Verifiable Rewards
AU  - Haechan Kim
AU  - Soohyun Ryu
AU  - Gyouk Chu
AU  - Doohyuk Jang
AU  - Eunho Yang
PY  - 2026
UR  - https://arxiv.org/abs/2603.18444
ID  - 2603.18444
ER  -