{
  "id": 3575,
  "url": "https://arxiv.org/abs/2605.28678v1",
  "title": "DREAM-R: Multimodal Speculative Reasoning with RL-Based Refined Drafting, Precise Verification, and Fully Parallel Execution",
  "summary": "Speculative reasoning has recently been proposed as a means to accelerate reasoning-intensive generation in large multimodal models, but its effectiveness is often constrained by misalignment between speculative drafts and target-verified reasoning. In this work, we introduce DREAM-R, a framework that substantially improves the performance of speculative reasoning. At its core, DREAM-R employs Speculative Alignment Policy Optimization (SAPO), a reinforcement-learning objective that trains draft ",
  "authors": "Yunhai Hu, Zining Liu, Xiangyang Yin, Tianhua Xia, Bo Bao, Eric Sather et al.",
  "category": "research",
  "topics": "regulation,safety-alignment",
  "orgs": null,
  "regions": null,
  "published_at": "2026-05-27T16:11:10.000Z",
  "fetched_at": "2026-07-14T16:30:23.244Z",
  "source_slug": "arxiv-ethics",
  "source_name": "arXiv",
  "source_homepage": "https://arxiv.org",
  "ethics_ai_record_url": "https://ethics.ai/record/3575",
  "original_url": "https://arxiv.org/abs/2605.28678v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}