Metraipolis › archive › dc_34de18d577ca

Simplex Relaxation for Discrete Diffusion — arXiv:2608.10615

Source: arXiv (en)
Original: https://arxiv.org/abs/2608.10615
License (as archived): CC0-1.0
Archived: 2026-10-19T06:14:13.463Z
SHA-256 of the archived text: 532c9ca3fddeffa141749aee03ba59b6801065e1404bfd2f9f993cbc1c92cfbb
Stable URL of this page: https://metraipolis.com/documento/dc_34de18d577ca

How to cite

Simplex Relaxation for Discrete Diffusion — arXiv:2608.10615 — arXiv (en), https://arxiv.org/abs/2608.10615 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_34de18d577ca (Metraipolis archive dc_34de18d577ca, sha256 532c9ca3fddeffa141749aee03ba59b6801065e1404bfd2f9f993cbc1c92cfbb)

BibTeX

@misc{dc34de18d577ca,
  title = {Simplex Relaxation for Discrete Diffusion — arXiv:2608.10615},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.10615},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_34de18d577ca; sha256 532c9ca3fddeffa141749aee03ba59b6801065e1404bfd2f9f993cbc1c92cfbb},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.10615. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Discrete diffusion models for categorical generation are defined by a corruption kernel, which determines the intermediate state space and the associated reverse prediction problem. Within this family, uniform diffusion has been extensively developed, with recent work connecting its categorical corruption process to continuous representations and dynamics. Motivated by this view, we ask whether uniform discrete diffusion can be augmented with an explicit continuous state while leaving its categorical corruption process unchanged. We introduce Simplax, an exact Dirichlet--categorical augmentation that couples each corrupted categorical state with an auxiliary simplex-valued variable while preserving the uniform diffusion process as its categorical marginal. This augmentation yields a tractable Rao--Blackwellized reverse-bridge objective and a stochastic reverse sampler, while retaining the corrupted categorical state as the denoiser input. Empirically, Simplax improves the generative perplexity--entropy tradeoff on unconditional OpenWebText generation. On Sudoku, a model trained exclusively on $30$-clue puzzles achieves the highest accuracy among the compared methods across all evaluated clue densities, including the minimum uniquely solvable $17$-clue regime, and also achieves the highest validity in unconditional generation.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-12 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2608.10615 — 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 2026-10-19T06:08:50.320Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-08-12?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).