ReSkill: Selective Self-Improvement for Robot Manipulation
ReSkill: Selective Self-Improvement for Robot Manipulation
This research presents ReSkill, a selective self-improvement framework for robot manipulation that identifies capability gaps using object-centric task graphs and iteratively revises stage organization, hand-off conditions, and recovery routes.
The paper introduces ReSkill, a framework that uses coding agents to iteratively revise robot manipulation policies by targeting specific capability gaps, validated on both simulated and real-world tasks with reported success rates of 86.2% on RoboDojo benchmarks and 77.5% across eight real-world trials.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
Reported numbers
median revision function change count
3
View original evidence
cceeds in 72 of 80 rebuilt-simulation trials and 62 of 80 real-robot trials (77.5%). The median revision changes only three functions, approximately 0.5% of a program, while 39% of task-specific functions in later-added tasks directly reuse existing implementations. These results show how locating local deficiencies, validating candidate changes, and accumulating reusable skills can improve execution of compositional
Open source S5
- paper: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
y. The agent participates only during development; the deployed policy calls the retained skills without invoking the agent at run time. On the same development set, average success across ten RoboDojo tasks rises from 30.0% for the initial versions to 86.2% for the final adopted programs; retained programs achieve 83.6% in recorded robustness and model-selection evaluations. Across eight real-world tasks, ReSkill su
Open source S4
cceeds in 72 of 80 rebuilt-simulation trials and 62 of 80 real-robot trials (77.5%). The median revision changes only three functions, approximately 0.5% of a program, while 39% of task-specific functions in later-added tasks directly reuse existing implementations. These results show how locating local deficiencies, validating candidate changes, and accumulating reusable skills can improve execution of compositional
Open source S5
Source：arXiv Robotics — research abstracts · arxiv.org