{"@context":"https://schema.org","@type":"CreativeWork","@id":"https://forgecascade.org/public/capsules/e7ac3a7a-993e-4dd1-beef-978457ab2201","identifier":"e7ac3a7a-993e-4dd1-beef-978457ab2201","url":"https://forgecascade.org/public/capsules/e7ac3a7a-993e-4dd1-beef-978457ab2201","name":"SURGE: Approximation-free Training Free Particle Filter for Diffusion Surrogate","text":"# SURGE: Approximation-free Training Free Particle Filter for Diffusion Surrogate\n\n**Authors:** Lifu Wei, Yinuo Ren, Naichen Shi, Yiping Lu\n**arXiv:** https://arxiv.org/abs/2605.18745v1\n**Published:** 2026-05-18T17:59:00Z\n\n## Abstract\nDiffusion-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 between path-wise and particle-wise SMC: the Girsanov path weight admits a backward conditional expectation that recovers the previous particle-level weights, guaranteeing that both schemes produce the same unbiased terminal law. Empirically, \\texttt{URGE} outperforms existing inference-time guidance baselines on synthetic tests and diffusion-model benchmarks, achieving better generation quality, while being significantly simpler to implement and fully gradient-free.","keywords":["stat.ML","cs.LG","math.NA","math.PR","q-fin.MF"],"about":[{"@type":"Thing","name":"Earth Lusca"},{"@type":"Thing","name":"APT37"},{"@type":"Thing","name":"Higaisa"},{"@type":"Thing","name":"Wingbird"},{"@type":"Thing","name":"Industroyer2"},{"@type":"Thing","name":"DropBook"},{"@type":"Thing","name":"vSphere Installation Bundles"},{"@type":"Thing","name":"Time Based Checks"},{"@type":"Thing","name":"User Activity Based Checks"}],"citation":[],"isPartOf":{"@type":"Dataset","name":"Forge Cascade Knowledge Graph","url":"https://forgecascade.org"},"publisher":{"@type":"Organization","name":"Forge Cascade","url":"https://forgecascade.org"},"dateCreated":"2026-05-19T06:00:07.233000Z","dateModified":"2026-05-19T06:00:07.233000Z","isBasedOn":"https://arxiv.org/abs/2605.18745v1","additionalProperty":[{"@type":"PropertyValue","name":"trust_level","value":65},{"@type":"PropertyValue","name":"verification_status","value":"source_linked"},{"@type":"PropertyValue","name":"evidence_level","value":"primary_source"}]}