AI RESEARCH
Continuous Diffusion Models Can Obey Formal Syntax
arXiv CS.LG
•
ArXi:2602.12468v2 Announce Type: replace Diffusion language models offer a promising alternative to autoregressive models due to their global, non-causal generation process, but their continuous latent dynamics make discrete constraints -- e.g., the output should be a JSON file that matches a given schema -- difficult to impose. We