{
  "id": "2611.06425",
  "title": "D3D: Decoupled Diffusion Drafting for Asynchronous Speculative Decoding",
  "authors": "Baoyou Chen, Ruiqiao Mei, Guancong Lin, Tao Song, Weihao Yuan, Qun Liu, Siyu Zhu",
  "first_public": "2026-11-04",
  "added_at": "2026-11-09",
  "venue": "arXiv",
  "arxiv": "2611.06425",
  "paper_url": "https://arxiv.org/abs/2611.06425",
  "html_url": "https://arxiv.org/html/2611.06425v1",
  "lens": [
    "beyond-sequential"
  ],
  "representation": [
    "discrete-token"
  ],
  "computation": [
    "diffusion",
    "parallel"
  ],
  "training": [
    "distillation",
    "supervised"
  ],
  "questions": [
    "efficiency",
    "architecture"
  ],
  "evidence": [
    "empirical"
  ],
  "modality": [
    "language"
  ],
  "summary": "Trains a standalone diffusion model to draft several tokens at once, then prunes it using verification acceptance as the criterion. Prepares continuation branches while the target model checks the current draft, reusing a branch when its prefix matches.",
  "provenance": "Primary-source literature review, 2026-11-09",
  "author_names": [
    "Baoyou Chen",
    "Ruiqiao Mei",
    "Guancong Lin",
    "Tao Song",
    "Weihao Yuan",
    "Qun Liu",
    "Siyu Zhu"
  ],
  "record_url": "https://latent-implicit-thinking.github.io/papers/2611.06425/",
  "legacy_record_url": "https://latent-implicit-thinking.github.io/papers/#paper-2611-06425",
  "citation_date": "2026-11-04",
  "json_url": "https://latent-implicit-thinking.github.io/papers/2611.06425/record.json",
  "markdown_url": "https://latent-implicit-thinking.github.io/papers/2611.06425/index.md",
  "bibtex_url": "https://latent-implicit-thinking.github.io/papers/2611.06425/citation.bib",
  "csl_url": "https://latent-implicit-thinking.github.io/papers/2611.06425/citation.csl.json",
  "pdf_url": "https://arxiv.org/pdf/2611.06425"
}