TY  - RPRT
TI  - Maximizing Rollout Informativeness under a Fixed Budget: A Submodular View of Tree Search for Tool-Use Agentic Reinforcement Learning
AU  - Yuelin Hu
AU  - Zhenbo Yu
AU  - Zhengxue Cheng
AU  - Wei Liu
AU  - Li Song
PY  - 2026
UR  - https://arxiv.org/abs/2605.05262
ID  - 2605.05262
ER  -