TRACE 高效代理强化学习的统一预算分配框架 TRACE A Unified Rollout Budget Allocation Framework for Efficient Agentic Reinforcement Learning.docx

TRACE 高效代理强化学习的统一预算分配框架 TRACE A Unified Rollout Budget Allocation Framework for Efficient Agentic Reinforcement Learning.docx

2026-06-10

1

TRACE:AUnifiedRolloutBudgetAllocationFrameworkforEfficient

AgenticReinforcementLearning

HemingZou1,2YunQu1,2,LizhouCai1,YixiuMao1,

,WeijieLiu2,KaiYang2,SaiyongYang2XiangyangJi1?

1TsinghuaUniversity2LLMDepartment,Tencent

囡cheemswang@

stevesyang@

xyj

您可能关注的文档

文档评论(0)

1亿VIP精品文档

相关文档