{
  "id": 16157,
  "url": "https://arxiv.org/abs/2608.01263v1",
  "title": "Distill What the Student Can See: Fisher-Projected On-Policy Distillation for Vision-Language Models",
  "summary": "On-policy distillation (OPD) samples trajectories from the current student policy and minimizes token-level divergence between student and teacher next-token distributions at prefixes along those trajectories. This aligns the distillation states with the student's own generation distribution. However, it still assumes that the complete teacher distribution is an appropriate target across student capacities. In vision--language reasoning, teacher corrections can depend on visual distinctions that",
  "authors": "Leyan Xue, Feng Xiong, Mingjun Ma, Changqing Zhang",
  "category": "research",
  "topics": "regulation,children-education",
  "orgs": null,
  "regions": null,
  "published_at": "2026-08-02T14:16:14.000Z",
  "fetched_at": "2026-08-04T05:10:21.797Z",
  "source_slug": "arxiv-cslg",
  "source_name": "arXiv cs.LG",
  "source_homepage": "https://arxiv.org/list/cs.LG/recent",
  "ethics_ai_record_url": "https://ethics.ai/record/16157",
  "original_url": "https://arxiv.org/abs/2608.01263v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}