AndroidWorld
Emerging15papers using it
2025first seen
AndroidWorld is a dataset/benchmark used to evaluate the performance of Android agents in reinforcement learning tasks.
Papers using AndroidWorld (15)
- UI-S1: Advancing GUI Automation via Semi-online Reinforcement LearningHi-Agent: Hierarchical Vision-Language Agents for Mobile Device ControlMobileRL: Online Agentic Reinforcement Learning for Mobile GUI AgentsAndroid Coach: Improve Online Agentic Training Efficiency with Single State Multiple ActionsK^2-Agent: Co-Evolving Know-What and Know-How for Hierarchical Mobile Device ControlOS-Themis: A Scalable Critic Framework for Generalist GUI RewardsAdaptive Milestone Reward for GUI AgentsOS-Oracle: A Comprehensive Framework for Cross-Platform GUI Critic ModelsSTEP: Success-Rate-Aware Trajectory-Efficient Policy OptimizationGRACE: A Language Model Framework for Explainable Inverse Reinforcement LearningDyna-Mind: Learning to Simulate from Experience for Better AI AgentsSucceed or Learn Slowly: Sample Efficient Off-Policy Reinforcement Learning for Mobile App ControlUI-TARS-2 Technical Report: Advancing GUI Agent with Multi-Turn Reinforcement LearningGUI-Shepherd: Reliable Process Reward and Verification for Long-Sequence GUI TasksMobile-Agent-v3: Fundamental Agents for GUI Automation