Preprint
Aug 2026
Physics-Guided Generative Surrogates for Parametric Rarefied Flows with Neural-Field Auto-Decoders: A Pipeline-Level Study of Flow Matching and Diffusion
The results validate sampling-based conditional transport on the shared representation as an effective steady surrogate, with a natural route to multivalued or stochastic solution families, with a natural route to multivalued or stochastic solution families.
Yilun Qi, Guan Zhang, Xu Wang et al.
· 0 citations