analiza aiqbench
This research addresses a critical bottleneck in robotic manipulation by demonstrating how targeted, dense data collection specifically for contact-rich segments can drastically improve success rates without relying on teleoperators or continuous online updates. By combining this focused dataset strategy with offline deep reinforcement learning and traditional planning for free-space motion, the approach achieves high generalization even in out-of-distribution scenarios where standard end-to-end models fail. This work advances machine learning applications in robotics by proving that strategic data efficiency yields superior performance over brute-force training on large datasets.
Data and Learning Where it Matters for Contact-Rich Manipulation
arXiv:2607.15982v1 Announce Type: new Abstract: Learned policies trained end-to-end on large datasets often remain brittle in high-precision tasks and struggle with generalization. We find that these limitations largely stem from a lack of structure and focus in data collection. Our key insight is to leverage dense data collection only for the critical segment of contact-rich tasks and to rely on traditional planning during simple free-space motion. We propose an automated data-collection scheme in combination with offline deep reinforcement learning for the critical segment of the task, eliminating reliance on a teleoperator's skill and on online policy updates. Across four challenging real-world tasks, using only 2 to 2.5 hours of autonomous data collection, we achieve an average success rate of 96%, compared to the strongest baseline at 55%. Notably, performance remains high in out-of-distribution scenarios where end-to-end approaches struggle. Our results pave the way for targeted data collection for contact-rich tasks and for high success rates in precision applications.
Abstract:Learned policies trained end-to-end on large datasets often remain brittle in high-precision tasks and struggle with generalization. We find that these limitations largely stem from a lack of structure and focus in data collection. Our key insight is to leverage dense data collection only for the critical segment of contact-rich tasks and to rely on traditional planning during simple free-space motion. We propose an automated data-collection scheme in combination with offline deep reinforcement learning for the critical segment of the task, eliminating reliance on a teleoperator's skill and on online policy updates. Across four challenging real-world tasks, using only 2 to 2.5 hours of autonomous data collection, we achieve an average success rate of 96%, compared to the strongest baseline at 55%. Notably, performance remains high in out-of-distribution scenarios where end-to-end approaches struggle. Our results pave the way for targeted data collection for contact-rich tasks and for high success rates in precision applications.
Robotics (cs.RO)
arXiv:2607.15982 [cs.RO]
(or arXiv:2607.15982v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2607.15982 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 246 cuvinte
Subiecte conexe
The State of Simulation for Physical AI: An Overview
Hugging Face · 21 iul. 2026
Building a Foundation Stack for General-Purpose Robots
IEEE Spectrum · 13 iul. 2026
Play2Perfect: What Matters in Dexterous Play Pretraining for Precise Assembly?
arXiv · 9 sept. 2026
ContactWorld: What Matters in Vision-Tactile World Models for Contact-Rich Manipulation
arXiv · 16 iun. 2026