[2603.21342] Generalized Discrete Diffusion from Snapshots

Statistics > Machine Learning arXiv:2603.21342 (stat) [Submitted on 22 Mar 2026] Title:Generalized Discrete Diffusion from Snapshots Authors:Oussama Zekri, Théo Uscidda, Nicolas Boullé, Anna Korba View a PDF of the paper titled Generalized Discrete Diffusion from Snapshots, by Oussama Zekri and 3 other authors View PDF HTML (experimental) Abstract:We introduce Generalized Discrete Diffusion from Snapshots (GDDS), a unified framework for discrete diffusion modeling that supports arbitrary noising processes over large discrete state spaces. Our formulation encompasses all existing discrete diffusion approaches, while allowing significantly greater flexibility in the choice of corruption dynamics. The forward noising process relies on uniformization and enables fast arbitrary corruption. For the reverse process, we derive a simple evidence lower bound (ELBO) based on snapshot latents, instead of the entire noising path, that allows efficient training of standard generative modeling architectures with clear probabilistic interpretation. Our experiments on large-vocabulary discrete generation tasks suggest that the proposed framework outperforms existing discrete diffusion methods in terms of training efficiency and generation quality, and beats autoregressive models for the first time at this scale. We provide the code along with a blog post on the project page : \href{this https URL}{this https URL}. Comments: Subjects: Machine Learning (stat.ML); Artificial Intelligence (cs....