news

Feb 24, 2026 🎉 Excited to share our latest work, ACuRL — an Autonomous Curriculum Reinforcement Learning framework that enables fully autonomous continual learning for computer-use agents, with zero human data.
Feb 24, 2026 🎉 Online-Mind2Web has been accepted to COLM 2025! See you in Montreal!🍁
Oct 09, 2025 🎉 Excited to share our latest work “Early Experience”, where agents learn from their own experience via implicit world modeling and self-reflection — a reward-free and scalable paradigm.