Metraipolis › archive › dc_537bf322504e

Learning Graph Energies for Finite-Time Generation — arXiv:2610.19946

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.19946
License (as archived)
CC0-1.0
Archived
2026-10-15T14:18:46.277Z
SHA-256 of the archived text
f2e2ec90706112ca9d8ffdaa633f989ae1a0aa903d42d351a142f7e3ee9284cf
Stable URL of this page
https://metraipolis.com/documento/dc_537bf322504e

How to cite

Learning Graph Energies for Finite-Time Generation — arXiv:2610.19946 — arXiv (en), https://arxiv.org/abs/2610.19946 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_537bf322504e (Metraipolis archive dc_537bf322504e, sha256 f2e2ec90706112ca9d8ffdaa633f989ae1a0aa903d42d351a142f7e3ee9284cf)
BibTeX
@misc{dc537bf322504e,
  title = {Learning Graph Energies for Finite-Time Generation — arXiv:2610.19946},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.19946},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_537bf322504e; sha256 f2e2ec90706112ca9d8ffdaa633f989ae1a0aa903d42d351a142f7e3ee9284cf},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.19946. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Energy-based models can represent distributions over discrete structures such as molecular graphs, but an energy that is accurate at equilibrium need not yield accurate samples when the sampler stops after a fixed number of steps. We introduce finite-horizon energy learning, which adapts the acceptance energy of a Metropolis--Hastings sampler to the distribution it returns after a prescribed number of steps, with initialization, proposals, and annealing schedule held fixed. For graph generation, a frozen pretrained model supplies initialization and proposals, and a learned energy governs acceptance. A training-only discriminator scores completed samples, and coupled continuations of the accepted and rejected branches of a proposal give an acceptance-gradient estimator that is unbiased for a fixed discriminator under stated conditions, without equilibrated negative samples or differentiation through discrete states. Generation uses the learned energy in the original single-trajectory sampler, without discriminator evaluations or reranking. On twelve pairwise binary targets with exactly evaluated terminal laws, learned energies have lower median terminal total variation than the equilibrium-correct energy in every tested short-horizon setting, and than its numerically tuned scalar multiple in most settings. On MOSES, a learned two-coefficient combination of pretrained Graph Energy Matching (GEM) potentials lowers FCD by 3.7--10.5\% relative to scale-calibrated GEM at matched sampling steps, with validity gains at the shortest budget and a sub-point validity trade-off at higher budgets.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.19946 — 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 2026-10-15T13:53:35.679Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).