ManiFlow：一种基于一致性流训练的通用机器人操作策略
ManiFlow: A General Robot Manipulation Policy via Consistency Flow Training
AI总结：
本文提出通用机器人操作策略ManiFlow，结合流匹配与一致性训练实现少步推理生成高维灵巧动作，通过DiT-X架构处理多模态输入，在仿真与真实多类型机器人任务中大幅提升成功率，兼具强泛化性与数据扩展能力。
摘要：
本文提出ManiFlow，一种面向通用机器人操作的视觉运动模仿学习策略，可基于多样的视觉、语言和本体感觉输入生成精确的高维动作。我们结合流匹配与一致性训练，仅需1-2步推理即可生成高质量的灵巧动作。为高效处理多样输入模态，我们提出DiT-X——一种带自适应交叉注意力和AdaLN-Zero条件的扩散Transformer架构，可实现动作token与多模态观测间的细粒度特征交互。ManiFlow在各类仿真基准上表现出持续提升，在单臂、双臂及仿人机器人等灵巧度递增的真实任务中，成功率近乎翻倍。大量评估进一步证明ManiFlow对新物体和背景变化具有强鲁棒性与泛化性，且在更大规模数据集下展现出优异的扩展能力。项目主页：maniflow-policy.github.io。
英文摘要：
This paper introduces ManiFlow, a visuomotor imitation learning policy for general robot manipulation that generates precise, high-dimensional actions conditioned on diverse visual, language and proprioceptive inputs. We leverage flow matching with consistency training to enable high-quality dexterous action generation in just 1-2 inference steps. To handle diverse input modalities efficiently, we propose DiT-X, a diffusion transformer architecture with adaptive cross-attention and AdaLN-Zero conditioning that enables fine-grained feature interactions between action tokens and multi-modal observations. ManiFlow demonstrates consistent improvements across diverse simulation benchmarks and nearly doubles success rates on real-world tasks across single-arm, bimanual, and humanoid robot setups with increasing dexterity. The extensive evaluation further demonstrates the strong robustness and generalizability of ManiFlow to novel objects and background changes, and highlights its strong scaling capability with larger-scale datasets. Our website: maniflow-policy.github.io.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
