AI RESEARCH
Latent-Augmented Discrete Diffusion Models
arXiv CS.AI
•
ArXi:2510.18114v3 Announce Type: replace-cross Discrete diffusion models have emerged as a powerful class of models and a promising route to fast language generation, but practical implementations typically rely on factored reverse transitions ignoring cross-token dependencies and degrading few-step performance. We propose Latent-Augmented Discrete Diffusion (LADD), which