MMODELYST
People/Yao Hu
PPL

Yao Hu

Papers (3)
Evolving-RL: End-to-End Optimization of Experience-Driven Self-Evolving Capability within AgentsHTPO: Towards Exploration-Exploitation Balanced Policy Optimization via Hierarchical Token-level Objective ControlKnowledge-Graph Paths as Intermediate Supervision for Self-Evolving Search Agents