AI RESEARCH
VISD: Enhancing Video Reasoning via Structured Self-Distillation
arXiv CS.CV
•
ArXi:2605.06094v1 Announce Type: new