Human Demonstrations as Post-Training Data for Any Robot Bod
Overview
Source roundup from published reports. Claims below are attributed to their publishers, not independently verified. arXiv Robotics — research abstracts: Human Demonstrations as Post-Training Data for Any Robot Body: Retargeting the Object, Not the Hand. This research explores using human demonstrations as post-training data for robots with different body structures. By retargeting object motion rather than hand motion, the method enables scalable deployment across diver…
Generated from attributed reports · Updated 2 hours ago
Event evidence and corrections
0 attributed source owners. Ownership does not establish independent confirmation. Quantities are reported separately and are never added together.
Environment: simulation
“bot hands and objects from these simulated trajectories to produce physically faithful observations synchronized with state and control records. Rendering the same trajectories in diverse, task-compatible backgrounds expands paired training data while preserving robot-object motion and contact. Across TACO, OakInk2, and self-collected data, EgoMatrix achieves 54.84% end-to-end success in simulation versus 14.52% for”
Exact source · revision 2Source owner not reported
Reported quantity · captures: 42 other · Basis not reported
“ion against the demonstrated object motion as ground truth: the human-to-simulation leg of human-to-simulation-to-robot transfer, measured per robot. Each (robot, episode) pair gets one outcome: pass, failure with a named mode, or refusal, with its reason, before any motion. The method runs as a working system over a growing collection; the set analysed here is a sample of it: 42 captures solved on 8 robots, 157 pair”
Exact source · revision 1Source owner not reported
Reported quantity · robot pairs: 157 other · Basis not reported
“ion against the demonstrated object motion as ground truth: the human-to-simulation leg of human-to-simulation-to-robot transfer, measured per robot. Each (robot, episode) pair gets one outcome: pass, failure with a named mode, or refusal, with its reason, before any motion. The method runs as a working system over a growing collection; the set analysed here is a sample of it: 42 captures solved on 8 robots, 157 pair”
Exact source · revision 1Source owner not reported
Reported quantity · refusals that failed: 13 other · Basis not reported
“s, of which 31.2% pass, 52.2% fail, 14.0% are refused and 2.5% are unauditable (kinematic feasibility, +/-50 mm / +/-15 deg, in simulation). Refusals held up: every pair the checks refuse on stored evidence yet ran anyway failed (13 of 13; 5 on a body whose public model welds its hand). A description file could not be shown to screen a robot: its outcomes were barely more predictable than guessing. No policy has yet”
Exact source · revision 1Source owner not reported
Reported quantity · refusals on a body with a public model: 5 other · Basis not reported
“s, of which 31.2% pass, 52.2% fail, 14.0% are refused and 2.5% are unauditable (kinematic feasibility, +/-50 mm / +/-15 deg, in simulation). Refusals held up: every pair the checks refuse on stored evidence yet ran anyway failed (13 of 13; 5 on a body whose public model welds its hand). A description file could not be shown to screen a robot: its outcomes were barely more predictable than guessing. No policy has yet”
Exact source · revision 1Source owner not reported
Report timeline
Follow attributed reports and material updates.
- arXiv Robotics — research abstractsCross-Embodiment Causal Transfer from Human Videos to Robot Manipulation
This research explores how to transfer knowledge from human manipulation videos to robot manipulation by analyzing subtask transfer through causal beams. The proposed method, Causal-Beam-Guided Subtask Transfer (CBST), learns manipulation-related representations from human videos and grounds them in task-level semantic variables using robot demonstrations. Experiments show improvements in task success over baseline world-action models.
- arXiv Robotics — research abstractsHuman Demonstrations as Post-Training Data for Any Robot Body: Retargeting the Object, Not the Hand
This research explores using human demonstrations as post-training data for robots with different body structures. By retargeting object motion rather than hand motion, the method enables scalable deployment across diverse robotic embodiments. The study evaluates the approach on 42 captures across 8 robots, showing varying success rates and highlighting challenges in sim-to-real transfer.
- arXiv Robotics — research abstractsEgoMatrix: Physically Faithful Dexterous Robot Demonstrations from Egocentric Human Videos
EgoMatrix is an end-to-end data engine that converts monocular egocentric human videos into bimanual dexterous trajectories and paired robot observation videos. The method focuses on relative geometry and contact reconstruction, and uses reinforcement learning to reproduce tasks in simulation.
Timezone · UTC
Article dates follow your selected timezone. Briefing editions use Hong Kong time (UTC+8).