arxivcs.RO2026-06-26
WARP-RM: A Warp-Augmented Relative Progress Reward Model for Data Curation
Justin Yu, Andrew Goldberg, Kavish Kondap, Karim El-Refai, Ethan Ransing, Qianzhong Chen, et al.
Scaling imitation learning requires large datasets, yet human teleoperation inevitably produces mixed-quality demonstrations containing hesitations and recoveries. Prior frame-level progress reward models supervise on absolute temporal progress proxies that suffer from label nois…