Lookahead Sample Reward Guidance for Test-Time Scaling of Diffusion Models 文章

ArXiv CS.AI2026-06-02NEWSen作者: Yeongmin Kim, Donghyeok Shin, Byeonghu Na, Minsang Park, Richard Lee Kim, Il-Chul Moon

查看原文 →

Lookahead Sample Reward Guidance for Test-Time Scaling of Diffusion Models · 相关技术

相关技术

Expected future reward accurate solver few-step lookahead sampling closed-form reward guidance neural backpropagation gradient guidance sequential Monte Carlo Tweedie-based approaches backward rollout diffusion models