AI RESEARCH
Unrestrained Simplex Denoising for Discrete Data. A Non-Markovian Approach Applied to Graph Generation
arXiv CS.LG
•
ArXi:2603.28572v1 Announce Type: new Denoising models such as Diffusion or Flow Matching have recently advanced generative modeling for discrete structures, yet most approaches either operate directly in the discrete state space, causing abrupt state changes. We