{
  "id": 19020,
  "url": "https://arxiv.org/abs/2608.12306v1",
  "title": "Redistribution-based Cost Inference Improves Sparse Safe Offline RL",
  "summary": "Safe offline RL typically assumes access to dense per-step cost annotations, but in practice supervisors provide only trajectory-level stop-feedback: a binary signal at the first unsafe transition, with no per-step attribution. We frame this as a temporal credit assignment problem and propose the Redistribution-based Cost Inference (RCI) framework, which converts sparse stop-feedback into dense per-step costs via return decomposition, then trains a constrained offline policy on the augmented dat",
  "authors": "Ebenezer Gelo, Geraud Nangue Tasse, Steven James, Benjamin Rosman",
  "category": "research",
  "topics": "regulation",
  "orgs": null,
  "regions": null,
  "published_at": "2026-08-12T17:53:15.000Z",
  "fetched_at": "2026-08-13T05:10:37.786Z",
  "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/19020",
  "original_url": "https://arxiv.org/abs/2608.12306v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}