Robots that Collaborate: Sequential Asymmetric Imitation for Learning Coupled Robot Policies
arXiv 论文《Robots that Collaborate: Sequential Asymmetric Imitation for Learning Coupled Robot Policies》发布 v3 修订版,研究协作式移动操作中机器人如何与只能部分观测的伙伴在物理交互中协调。同一批 arXiv 更新中还出现多篇相关投稿,涉及模仿策略对带噪物体位置的折扣学习、基于几何先验的快速非参数异方差模仿学习,以及 RecurIR 通过递归自我改进学习何时及如何修订修复策略。上述工作均以摘要形式公布,尚无进一步的实验或评审结论。
How this story unfolded
Key points6 items Model summary · ordered by when the evidence was published
- Point 01 11/16, 13:00
《Robots that Collaborate》更新至 v3,聚焦部分观测伙伴下的耦合机器人策略学习。
Based on 1 - Point 02 11/16, 13:00
该研究面向协作移动操作场景,要求机器人在物理交互中与伙伴协调。
Based on 1 - Point 03 11/16, 13:00
另一篇论文指出,机器人操作模仿策略常在带噪物体位置上训练,并学会折扣噪声读数。
Based on 1 - Point 04 11/16, 13:00
一项研究提出带几何先验的快速非参数异方差模仿学习,强调数据效率与快速适应新场景。
Based on 1 - Point 05 11/16, 13:00
RecurIR 通过递归自我改进,让图像修复智能体学习何时以及如何修订修复策略。
Based on 1 - Point 06 11/16, 13:00
RecurIR 针对异构、耦合的修复工具协调问题。
Based on 1
ReportsBy publish time · newer further down
- November 16, 2026 · Mon
- Start
RecurIR: Learning When and How to Revise Restoration Policies via Recursive Self-Improvement
Original summary arXiv:2611.18006v1 Announce Type: new Abstract: Agent-based image restoration has achieved promising performance by coordinating restoration tools to handle heterogeneous and coupl
- Start
Fast Non-Parametric Heteroscedastic Imitation Learning With Geometric Priors
Original summary arXiv:2610.08650v2 Announce Type: replace Abstract: When learning probabilistic policies from human demonstrations, data-efficient learning and fast adaptations to new scenarios ar
- Start
Robots that Collaborate: Sequential Asymmetric Imitation for Learning Coupled Robot Policies
Original summary arXiv:2606.16490v3 Announce Type: replace Abstract: Collaborative mobile manipulation requires robots to coordinate with a partially observed partner while physically interacting t
- Start
Imitation Policies Learn a Proprioceptive Anchor: A Predictive Law for Perception Noise and Its Closed Loop Cost
Original summary arXiv:2611.17658v1 Announce Type: new Abstract: Imitation policies for robot manipulation are often trained on noisy object positions, and they learn to discount the noisy reading.
arXiv cs.RO Original ↗ - Still developing · later reports will keep extending this axis