FlowRefiner: Flow Matching-Based Iterative Refinement for 3D Turbulent Flow Simulation

Yilong Dai Yiming Sun Yiheng Chen Shengyu Chen Xiaowei Jia Runlong Yu · 2026

Read FlowRefiner: Flow Matching-Based Iterative Refinement for 3D Turbulent Flow Simulation by Yilong Dai Yiming Sun Yiheng Chen Shengyu Chen Xiaowei Jia Runlong Yu (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-17149.txt.

Description

Dai and five colleagues propose FlowRefiner, a refiner for 3D turbulent flows. It swaps noisy steps for ODE correction, uses one training goal across all stages, and fixes the noise range at any depth via a sigma schedule. On large-scale 3D turbulence, it achieves top autoregressive accuracy with strong physics consistency.

Catalog record

CREATORYilong Dai Yiming Sun Yiheng Chen Shengyu Chen Xiaowei Jia Runlong Yu
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE57810
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.17149

Browse the full library · About the Tower of Babel