{
  "id": 18650,
  "url": "https://arxiv.org/abs/2608.11191v1",
  "title": "Test-Time Self-Evolving GUI Visual Grounding via Reflection-Guided On-Policy Self-Distillation",
  "summary": "GUI Visual Grounding is a fundamental capability for GUI agents. Existing models typically freeze their parameters after deployment, limiting their ability to adapt to unseen interfaces. Although recent methods attempt to adapt models via test-time reinforcement learning, they cannot reflect upon failed exploration. To overcome this, we propose a Test-Time Self-Evolving framework that enables models to improve after deployment without human-annotated ground truth. It constructs a closed-loop of",
  "authors": "Shiyu Xuan, Zechao Li",
  "category": "research",
  "topics": "regulation,agents-autonomy",
  "orgs": null,
  "regions": null,
  "published_at": "2026-08-11T17:50:25.000Z",
  "fetched_at": "2026-08-12T05:10:43.828Z",
  "source_slug": "x-arxiv-cs-ai",
  "source_name": "arXiv cs.AI",
  "source_homepage": "https://arxiv.org/list/cs.AI/recent",
  "ethics_ai_record_url": "https://ethics.ai/record/18650",
  "original_url": "https://arxiv.org/abs/2608.11191v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}