Metraipolis › archive › dc_8327fa5d6837

Grab a Coffee: Future-Aware Guidance for Discrete Diffusion with Compiled Objectives — arXiv:2609.35924

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.35924
License (as archived)
CC0-1.0
Archived
2026-10-07T16:22:07.770Z
SHA-256 of the archived text
62fb29c9884e84f23a07777c0d3111411363107a8994a24ddb54a8f601bebd7e
Stable URL of this page
https://metraipolis.com/documento/dc_8327fa5d6837

How to cite

Grab a Coffee: Future-Aware Guidance for Discrete Diffusion with Compiled Objectives — arXiv:2609.35924 — arXiv (en), https://arxiv.org/abs/2609.35924 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_8327fa5d6837 (Metraipolis archive dc_8327fa5d6837, sha256 62fb29c9884e84f23a07777c0d3111411363107a8994a24ddb54a8f601bebd7e)
BibTeX
@misc{dc8327fa5d6837,
  title = {Grab a Coffee: Future-Aware Guidance for Discrete Diffusion with Compiled Objectives — arXiv:2609.35924},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.35924},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_8327fa5d6837; sha256 62fb29c9884e84f23a07777c0d3111411363107a8994a24ddb54a8f601bebd7e},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.35924. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Discrete diffusion models generate sequences by iteratively resolving multiple tokens in parallel, offering a flexible alternative to left-to-right generation. However, guiding this process with a sequence-level objective is difficult because the value of one unresolved token depends on the other tokens with which it can form a high-reward sequence. Enumerating all such completions makes the whole guidance computation grow exponentially with the number of unresolved positions. We introduce COFFEE, a plug-and-play framework that avoids this enumeration by separating sequence dependence from the objective. At each diffusion step, a target-free carrier absorbs the marginal token distributions predicted by the denoiser to construct a joint model over the unresolved tokens, while a compiled finite-state model records how their combinations affect the sequence-level preference. Pairing their states allows COFFEE to transfer global preferences to unresolved positions and sample a clean reconstruction without retraining the diffusion model. The same framework supports explicit hard constraints and learned soft objectives. We evaluate COFFEE across multiple symbolic, language, and biological benchmarks, where it achieves strong control results with task-dependent quality and diversity trade-offs. By making objectives available to inference rather than only evaluation, COFFEE brings joint conditioning, completion-weighted guidance, and optimization-based constraints into pretrained neural generation, showing the potential of neural-symbolic methods in diffusion guidance.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-30 in cs.AI).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.35924 — 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 2026-10-07T15:54:53.448Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-30?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).