← 研踪 PaperTrailarXiv:2611.23929 · v1
ProxySearch: Decoupled Inference-Time Scaling for Diffusion Models via Asymmetric Noise-Rank Transfer
首次提交 2026-11-18 14:20
版本更新 2026-11-18 14:20
雷达观测 2026-11-19 19:08
摘要
Noise search at inference time improves text-to-image diffusion but requires many multi-step teacher rollouts per query. Distilled few-step models are now routinely released, yet their potential as proxies inside noise search is uncharacterized. We measure proxy-teacher rank agreement across three distillation regimes and three backbones (DiT, MMDiT, UNet) and find a sharp asymmetry. On our main pool, closed-form image statistics transfer from the proxy to the teacher (Spearman up to 0.80), while learned semantic scorers do not ( 0.12). Paired null controls trace this rank to the first denoising step for coarse statistics and to distillation for finer ones. We use the asymmetry for noise search under a joint constraint: a semantic reward specifies content, and a numeric target specifies image statistics. In PROXYSEARCH, the proxy filters candidates by statistical match in 1 step, and the teacher returns the survivor with the highest semantic reward. No weight trades one objective against the other. On FLUX-1024, PROXYSEARCH tightens the statistical target by 21-23% at the same number of teacher trajectories, with no detected loss on the optimized reward.
作者备注:Accepted to NeurIPS 2026. 26 pages, 7 figures
观测到的版本
相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。
v1 · 2026-11-18 14:20 · 最新观测版本
ProxySearch: Decoupled Inference-Time Scaling for Diffusion Models via Asymmetric Noise-Rank Transfer
Noise search at inference time improves text-to-image diffusion but requires many multi-step teacher rollouts per query. Distilled few-step models are now routinely released, yet their potential as proxies inside noise search is uncharacterized. We measure proxy-teacher rank agreement across three distillation regimes and three backbones (DiT, MMDiT, UNet) and find a sharp asymmetry. On our main pool, closed-form image statistics transfer from the proxy to the teacher (Spearman up to 0.80), while learned semantic scorers do not ( 0.12). Paired null controls trace this rank to the first denoising step for coarse statistics and to distillation for finer ones. We use the asymmetry for noise search under a joint constraint: a semantic reward specifies content, and a numeric target specifies image statistics. In PROXYSEARCH, the proxy filters candidates by statistical match in 1 step, and the teacher returns the survivor with the highest semantic reward. No weight trades one objective against the other. On FLUX-1024, PROXYSEARCH tightens the statistical target by 21-23% at the same number of teacher trajectories, with no detected loss on the optimized reward.
查看此版本解读 →查看此版本原文 ↗