High-Fidelity One-Step Generative Visuomotor Policy via Recursive Correction, Frequency Consistency, and Contrastive Flow Matching
通过递归校正、频率一致性和对比流匹配实现高保真一步生成视觉运动策略
机构 * School of Safety Science and Engineering, Anhui University of Science and Technology(安徽理工大学安全科学与工程学院) ; State Key Laboratory of Digital Intelligent Technology for Unmanned Coal Mining, Anhui University of Science and Technology(安徽理工大学煤矿智能开采技术与装备国家重点实验室) ; School of Artificial Intelligence, Anhui University of Science and Technology(安徽理工大学人工智能学院)
专题命中 效率与蒸馏 :diffusion(abstract)
AI总结 研究针对生成模型多步采样有延迟、一步加速方法有偏差等问题,提出含递归校正、频率一致性和对比流匹配三机制的框架,解决相关问题,实验显示该方法在低延迟下性能优。