When Do Local Score Models Extrapolate Across Size? A Diagnostic Theory and Benchmark
Researchers have developed a new diagnostic theory and benchmark to understand how well local score models can extrapolate across different system sizes. They found that architectural locality alone is insufficient for stable size extrapolation, which is instead governed by the quasi-locality of the Gaussian-smoothed score. The study introduces the Finite-Depth Local Flow (FDLF) benchmark to empirically validate these findings, demonstrating that stable extrapolation depends on the interplay between spatial mixing, score quasi-locality, and model receptive fields. AI
IMPACT Provides a theoretical framework and diagnostic tool to improve the reliability of AI models in scientific generative modeling tasks.