Skip to main content
Search

# Motion-Aware Reward Model

1 related articles

Subscribe to updates
  1. arXiv

    Video-editing rewards mistake frozen clips for quality: an inter-frame dynamics input cuts VE-Bench motion-ladder capture from 97% to 0%

    Instruction-guided video-editing reward models treat temporal consistency as quality and so favor candidates whose motion has been suppressed: on the VE-Bench motion ladder, the evaluated composite reward picked a motion-suppressed candidate in 97% of cases. A motion-aware reward model feeds per-frame RGB differences in as an explicit input and trains on motion-suppressed and fabricated-motion negatives, with a distillation anchor preserving alignment to human quality ratings; it cuts ladder capture to 0% and also detects segment freeze, burst freeze and flow attenuation that were absent from training.

You’re all caught up