Metraipolis › archive › dc_0bc95dc5efe9

Let It Go or Learn to Self-Correct: Continuous Diffusion for Constrained Discrete Tasks — arXiv:2609.09009

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.09009
License (as archived)
CC0-1.0
Archived
2026-10-16T13:58:27.695Z
SHA-256 of the archived text
ca709351a62376c26d6d94c50cf04d728563c26152ca06bdd298212c7b473bb7
Stable URL of this page
https://metraipolis.com/documento/dc_0bc95dc5efe9

How to cite

Let It Go or Learn to Self-Correct: Continuous Diffusion for Constrained Discrete Tasks — arXiv:2609.09009 — arXiv (en), https://arxiv.org/abs/2609.09009 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_0bc95dc5efe9 (Metraipolis archive dc_0bc95dc5efe9, sha256 ca709351a62376c26d6d94c50cf04d728563c26152ca06bdd298212c7b473bb7)
BibTeX
@misc{dc0bc95dc5efe9,
  title = {Let It Go or Learn to Self-Correct: Continuous Diffusion for Constrained Discrete Tasks — arXiv:2609.09009},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.09009},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_0bc95dc5efe9; sha256 ca709351a62376c26d6d94c50cf04d728563c26152ca06bdd298212c7b473bb7},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.09009. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Denoising Diffusion Probabilistic Models (DDPMs) generate samples by starting from noise and repeatedly denoising while keeping each update close to the current noisy state. This behavior is effective in many continuous domains, but its role is less clear for globally constrained discrete tasks, such as Sudoku, graph connectivity, Latin squares, and N-queens. In such settings, early discrete errors can be difficult to undo. As a result, standard diffusion sampling may preserve early mistakes, even when the model's clean predictions are informative. We compare standard samplers to sampling directly from the model's clean prediction. Without retraining, this single change improves Sudoku validity from 31% to 95%, with consistent gains across the other discrete tasks. We hypothesize that staying close to the current noisy state is harmful because the reverse trajectory can drift off the forward noising distribution the model was trained on. To reduce this train-test mismatch, we further introduce self-correction training, which exposes the model to its own predictions, improving robustness to errors that arise during inference. This substantially improves the performance of standard samplers. Our results suggest that continuous diffusion models can learn nontrivial global constraints, but discrete reasoning tasks require better alignment between training and inference: either through samplers that reduce commitment to early decisions, or through training that teaches the model to correct its own inference-time errors.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-09 in cs.LG).
Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.09009 — 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 2026-10-16T13:39:43.081Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-09?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).