AMAP-ML

AMAP-ML is the machine learning research organization behind LongHorizon-Harness, an open-source harness for running AI agents on long-horizon real-world tasks. The team, credited as the DreamX Team, publishes research on agent state management and verification, with contributors including Ziyu Ma, Hailang Huang, Shun Zou, Yong Wang, Shidong Yang, Yiming Hu, Fei Wei, and XiangXiang Chu. Their work focuses on improving how AI agents handle complex, multi-step workflows across GUI and CLI environments without context drift or unverified self-assessment.