[论文] SURGE: Approximation-free Training Free Particle Filter for Diffusion ...

论文概要

研究领域: ML 作者: Lifu Wei, Yinuo Ren, Naichen Shi 发布时间: 2026-05-19 arXiv: 2505.14304

中文摘要

基于扩散的生成模型日益依赖推理时引导，通过添加漂移项或重新加权专家混合来提升特定任务目标上的样本质量。然而，大多数现有技术需要重复的分数或梯度评估，引入了偏差、高计算开销或两者兼有。本文提出SURGE（通过Girsanov估计的无偏重采样），一种无导数的推理时缩放算法，通过Girsanov测度变换执行路径级重要性重加权。与先前工作计算基于梯度的粒子权重不同，SURGE为每条模拟轨迹附加一个简单的乘法权重并定期重采样。无需分数、无需Hessian、无需PDE评估。我们建立了路径级和粒子级SMC之间的等价性：Girsanov路径权重允许后向条件期望恢复先前的粒子级权重，保证两种方案产生相同的无偏终端分布。实证上，SURGE在合成测试和扩散模型基准上优于现有的推理时引导基线，实现更好的生成质量，同时实现显著更简单且完全无梯度。

原文摘要

Diffusion-based generative models increasingly rely on inference-time guidance, adding a drift term or reweighting mixture of experts, to improve sample quality on task-specific objectives. However, most existing techniques require repeated score or gradient evaluations, introducing bias, high computational overhead, or both. We introduce \texttt{URGE}, Unbiased Resampling via Girsanov Estimation, a derivative-free inference-time scaling algorithm that performs path-wise importance reweighting via a Girsanov change of measure. Instead of computing gradient-based particle weights in previous work, \texttt{URGE} attaches a simple multiplicative weight to each simulated trajectory and periodically resamples. No score, no Hessian, and no PDE evaluation is required. We establish an equivalence ...

自动采集于 2026-05-20

#论文 #arXiv #ML #小凯