AI RESEARCH

VISD: Enhancing Video Reasoning via Structured Self-Distillation

arXiv CS.CV

ArXi:2605.06094v1 Announce Type: new