Metraipolis › archive › dc_7993bbdd81ec

Dependency-Aware Revocable Decoding for Efficient Diffusion Large Language Model Inference — arXiv:2608.26574

Source: arXiv (en)
Original: https://arxiv.org/abs/2608.26574
License (as archived): CC0-1.0
Archived: 2026-10-17T15:19:34.089Z
SHA-256 of the archived text: 00d0f97f4dd528e4736dea4ed910eb7c25b3507bf368e0df00ab7fb9ed47ee9a
Stable URL of this page: https://metraipolis.com/documento/dc_7993bbdd81ec

How to cite

Dependency-Aware Revocable Decoding for Efficient Diffusion Large Language Model Inference — arXiv:2608.26574 — arXiv (en), https://arxiv.org/abs/2608.26574 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7993bbdd81ec (Metraipolis archive dc_7993bbdd81ec, sha256 00d0f97f4dd528e4736dea4ed910eb7c25b3507bf368e0df00ab7fb9ed47ee9a)

BibTeX

@misc{dc7993bbdd81ec,
  title = {Dependency-Aware Revocable Decoding for Efficient Diffusion Large Language Model Inference — arXiv:2608.26574},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.26574},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7993bbdd81ec; sha256 00d0f97f4dd528e4736dea4ed910eb7c25b3507bf368e0df00ab7fb9ed47ee9a},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.26574. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Diffusion large language models (dLLMs) offer a promising alternative to autoregressive generation by decoding multiple tokens in parallel through iterative denoising. However, increasing decoding parallelism often degrades generation quality, as early errors can contaminate later contexts. Revocable decoding mitigates this issue by re-evaluating decoded tokens and remasking unreliable ones, but existing methods overlook that unreliable tokens may also corrupt the verification context itself. We identify this failure mode and propose Dependency-Aware Revocable Decoding (DARD), a training-free framework that separates tokens into masked, candidate, and unmasked states. DARD verifies candidate tokens using a selective context that excludes less reliable tokens and adaptively regulates their influence on subsequent decoding. Experiments across 12 textual and multimodal benchmarks on 3 open-source dLLMs show that DARD consistently improves the speed-quality Pareto frontier over recent revocable decoding methods, achieving a 2.71$\times$ speedup and a 4.35-point CIDEr score gain over Saber on Flickr30K.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-28 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2608.26574 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-17T15:15:18.059Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-08-28?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).