Skip to content
Hot eventLive

Error-Corrected Inference-Time Scaling for Imperfect Diffusion Models

1 reports1 sources3 hr ago updated

Get the story

AI overview

2026-10-02,arXiv 发表一篇 Statistics Machine Learning 论文,提出 Energy-based Feynman-Kac Corrector(EBFKC),用于在推理阶段修正扩散模型的误差。方法利用 Feynman-Kac 动力学精确跟踪预设概率路径,并通过顺序蒙特卡洛与方差控制进行近似;同时以预训练能量作为代理,逐步消除终点失配。

Generated from reports · updated 3 hr ago

Timeline

Follow the coverage from different angles.

Oct 2, 2026
  1. arXiv · Statistics Machine Learning
    Error-Corrected Inference-Time Scaling for Imperfect Diffusion Models

    论文提出 Energy-based Feynman-Kac Corrector(EBFKC),在推理阶段通过参考能量实时修正扩散模型的误差。该方法利用 Feynman-Kac 动力学精确跟踪预设概率路径,并用顺序蒙特卡洛与方差控制引导进行近似,同时通过预训练能量作为代理逐步消除终点失配。

Heat trend

There is not enough continuous observation data to draw a trend yet.