Metraipolis › archive › dc_fd0fa3c52b6d
LatticeSMC: Where to Spend Inference-Time Compute in Chunked Sequence Generators — arXiv:2610.02774
- Source
- arXiv
- Original
- https://arxiv.org/abs/2610.02774
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-05T23:09:52.138Z
- SHA-256 of the archived text
77aa6d56f7bb0018e77ec93e79200b56b0bad4daeadc6f333d31025f7b880820- Stable URL of this page
https://metraipolis.com/documento/dc_fd0fa3c52b6d
How to cite
LatticeSMC: Where to Spend Inference-Time Compute in Chunked Sequence Generators — arXiv:2610.02774 — arXiv, https://arxiv.org/abs/2610.02774 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_fd0fa3c52b6d (Metraipolis archive dc_fd0fa3c52b6d, sha256 77aa6d56f7bb0018e77ec93e79200b56b0bad4daeadc6f333d31025f7b880820)
BibTeX
@misc{dcfd0fa3c52b6d,
title = {LatticeSMC: Where to Spend Inference-Time Compute in Chunked Sequence Generators — arXiv:2610.02774},
howpublished = {arXiv},
url = {https://arxiv.org/abs/2610.02774},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_fd0fa3c52b6d; sha256 77aa6d56f7bb0018e77ec93e79200b56b0bad4daeadc6f333d31025f7b880820},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv, https://arxiv.org/abs/2610.02774. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Long-form generators for music, motion and video produce sequences chunk by chunk, with each chunk generated by iterative denoising while rewards are defined over the full sequence. Existing inference-time steering methods typically act on one axis at a time: best-of-N at the end, Feynman-Kac steering across denoising steps, or streaming pruning across chunks, and are often compared under unmatched compute or different return rules. We introduce budget-matched chunked steering and propose LatticeSMC, a sampler derived from a Feynman-Kac model on the two-dimensional lattice of chunk index and denoising step. Two telescoping results make its design exact: for chunk-additive rewards, the two axes induce identical weights, so resampling should occur where lookahead is cheapest; for terminal rewards, any prefix score defines an exact intermediate potential, making prefix-evaluable rewards twists with no estimation or extra denoiser calls. LatticeSMC resamples on these potentials at chunk boundaries and, when scoring is free, within chunks, returning either a weighted draw or the best particle. Under matched compute, on music-to-dance diffusion and 40-second text-to-music generation, it raises beat alignment from 0.234 to 0.441 (best-of-N: 0.354) and prompt adherence from 0.470 to 0.560 at 32 particles, while preserving held-out quality. It also retains its advantage on long-range rewards and is preferred by human raters in 60-77 percent of pairwise comparisons. Finally, we show that commitment strength should follow the information in the current potential, while the value of lookahead is predicted by the within-set predictability of future reward. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno). Categorie arXiv: Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2610.02774 — 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 2026-10-05T23:01:38.633Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).