Problem
Research questions and friction points this paper is trying to address.
bimanual manipulation
human demonstration data
vision-language-action model
Scaling Bimanual Household Manipulation from 1,500 hours of Demonstrations to On-Policy Corrections
🤖 AI Summary
为解决双手机器人操作数据稀缺问题，研究通过1500小时演示训练VLA模型XR-2，并利用实时人类干预的数据进行优化。
📝 Abstract
Learning generalist policies for robust bimanual manipulation is bottlenecked by the scarcity of high quality large scale human demonstration data. In this work, we release 1,500 hours of diverse bimanual manipulation demonstrations covering everyday household tasks, and use this comprehensive corpus to train XR-2, a powerful vision-language-action (VLA) model. Enabled by a purpose built high throughput data pipeline and a carefully designed multi stage training paradigm, XR-2 attains strong manipulation performance in our systematic experiments while retaining favorable training efficiency and high data utilization. We further study two critical scaling axes: varying the amount of expert demonstration data, and post training on DAgger correction data from real time human interventions. In both settings, task success rate improves steadily over the data ranges we probe, exhibiting a clear consistent scaling trend at our current data scale. These results validate both the learning capacity of XR-2 and the promising scaling properties of the released dataset, which we open source to support reproducible research on bimanual robot manipulation learning.
Innovation
Methods, ideas, or system contributions that make the work stand out.
bimanual manipulation
vision-language-action model
large scale demonstration data
DAgger correction data
🔎 Similar Papers
No similar papers found.
