Robometer combines intra-trajectory progress supervision with inter-trajectory preference supervision on a 1M-trajectory dataset to learn more generalizable robotic reward functions than prior methods.
arXiv preprint arXiv:2306.10007 , year=
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
citation-role summary
background 2
citation-polarity summary
fields
cs.RO 3roles
background 2polarities
background 2representative citing papers
A low-cost whole-body teleoperation system enables effective imitation learning for complex bimanual mobile manipulation by co-training on mobile and static demonstration datasets.
A GPT-style model pre-trained on large video datasets achieves 94.9% success on CALVIN multi-task manipulation and 85.4% zero-shot generalization, outperforming prior baselines.
citing papers explorer
-
Robometer: Scaling General-Purpose Robotic Reward Models via Trajectory Comparisons
Robometer combines intra-trajectory progress supervision with inter-trajectory preference supervision on a 1M-trajectory dataset to learn more generalizable robotic reward functions than prior methods.
-
Mobile ALOHA: Learning Bimanual Mobile Manipulation with Low-Cost Whole-Body Teleoperation
A low-cost whole-body teleoperation system enables effective imitation learning for complex bimanual mobile manipulation by co-training on mobile and static demonstration datasets.
-
Unleashing Large-Scale Video Generative Pre-training for Visual Robot Manipulation
A GPT-style model pre-trained on large video datasets achieves 94.9% success on CALVIN multi-task manipulation and 85.4% zero-shot generalization, outperforming prior baselines.