PPL
Qi Gu
Papers (5)
GUI-CIDER: Mid-training GUI Agents via Causal Internalization and Density-aware Exemplar ReselectionSkill0.5: Joint Skill Internalization and Utilization for Out-of-Distribution Generalization in Agentic Reinforcement LearningWhen to Stop Reusing: Dynamic Gradient Gating for Sample-Efficient RLVRSelf-Distilled Agentic Reinforcement LearningSkill1: Unified Evolution of Skill-Augmented Agents via Reinforcement Learning