arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2606.09705cs.LGcond-mat.stat-mech
收藏

局部评分模型何时能跨尺寸外推?诊断理论与基准

When Do Local Score Models Extrapolate Across Size? A Diagnostic Theory and Benchmark

The University of Hong Kong(香港大学);Department of Physics and HK Institute of Quantum Science & Technology(物理系与香港量子科学与技术研究所)

Wenjie Xi

AI总结:

提出诊断理论,证明局部模型能否稳定外推取决于高斯平滑评分的准局部性,并引入有限深度局部流(FDLF)基准进行验证。

摘要:

科学生成建模通常需要尺寸迁移,即在小系统上训练的模型在大系统上评估。虽然平移不变架构允许这种评估,但我们表明架构局部性本身并不能保证稳定的尺寸外推。相反,稳定外推由高斯平滑评分的准局部性决定。通过Tweedie公式,远距离扰动可以通过后验协方差影响局部评分分量,这意味着局部模型只有在感受野覆盖平滑评分的响应范围时才能成功。我们形式化了这一机制,证明了反向扩散下局部边缘的尺寸一致比较定理。我们还引入了有限深度局部流(FDLF),这是一个具有精确评分、密度和可控响应范围的白盒诊断基准。实验上,我们验证了空间混合、平滑评分准局部性和模型感受野之间的相互作用。在空间混合下,平滑评分相对于感受野保持准局部性,从而实现稳定外推。相反,当空间混合减弱时,评分的局部性迅速退化,导致尺寸迁移失败。

英文摘要:

Scientific generative modeling often requires size transfer, where models trained on small systems are evaluated on larger ones. While translation-invariant architectures enable this evaluation, we show that architectural locality alone does not guarantee stable size extrapolation. Instead, stable extrapolation is governed by the quasi-locality of the Gaussian-smoothed score. Through Tweedie's formula, far-away perturbations can influence local score components via posterior covariance, meaning a local model succeeds only if its receptive field covers the smoothed score's response range. We formalize this mechanism, proving a size-uniform comparison theorem for local marginals under reverse diffusion. We also introduce Finite-Depth Local Flow (FDLF), a white-box diagnostic benchmark with exact scores, densities, and controllable response ranges. Empirically, we validate the interplay between spatial mixing, smoothed-score quasi-locality, and model receptive fields. Under spatial mixing, the smoothed score remains quasi-local relative to the receptive field, enabling stable extrapolation. Conversely, when spatial mixing weakens, the score's locality rapidly degrades, causing size transfer to fail.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑