aiqbench analysis
DexPIE addresses critical limitations in dexterous manipulation by introducing a post-training framework that refines policies using real-world deployment data rather than relying solely on demonstrations. By integrating asynchronous inference in relative action space and a continuous optimality indicator, the method reduces temporal noise and achieves a 37% improvement in success rates across complex tasks. This approach offers a robust pathway for scaling imitation learning in high-dimensional, contact-rich robotic environments.
DexPIE: Stable Dexterous Policy Improvement from Real-World Experience
arXiv:2606.09615v1 Announce Type: new Abstract: Dexterous manipulation presents substantial challenges for imitation learning due to its high-dimensional action space and complex contact-rich dynamics. Policies trained purely from demonstrations often suffer from compounding errors during deployment and require large amounts of expert data to achieve reliable performance. To move beyond the limitations of demonstration data, in this work, we propose DexPIE, a post-training framework for dexterous policy improvement from experience collected through real-world deployment. First, DexPIE enables effective exploration coverage through a dexterous-hand-adapted intervention system and multi-stage DAgger-style data collection across initial and intermediate task stages, providing reliable supervision for accurate policy evaluation. To reduce temporal noise between post-training rollouts and demonstration data, we introduce asynchronous inference in the relative action space, which better aligns rollout data with demonstrated behavior and allows the critic to learn a value function induced by a more consistent underlying policy. Finally, DexPIE improves the policy through conditioning on a continuous optimality indicator, allowing the policy to leverage the quality of data in a more fine-grained manner. Across three challenging real-world dexterous manipulation tasks, DexPIE achieves a 37% improvement in success rate over the demonstration-based reference policy, outperforming all baseline methods and demonstrating stronger robustness. The source code and dataset will be made publicly available.
Abstract:Dexterous manipulation presents substantial challenges for imitation learning due to its high-dimensional action space and complex contact-rich dynamics. Policies trained purely from demonstrations often suffer from compounding errors during deployment and require large amounts of expert data to achieve reliable performance. To move beyond the limitations of demonstration data, in this work, we propose DexPIE, a post-training framework for dexterous policy improvement from experience collected through real-world deployment. First, DexPIE enables effective exploration coverage through a dexterous-hand-adapted intervention system and multi-stage DAgger-style data collection across initial and intermediate task stages, providing reliable supervision for accurate policy evaluation. To reduce temporal noise between post-training rollouts and demonstration data, we introduce asynchronous inference in the relative action space, which better aligns rollout data with demonstrated behavior and allows the critic to learn a value function induced by a more consistent underlying policy. Finally, DexPIE improves the policy through conditioning on a continuous optimality indicator, allowing the policy to leverage the quality of data in a more fine-grained manner. Across three challenging real-world dexterous manipulation tasks, DexPIE achieves a 37% improvement in success rate over the demonstration-based reference policy, outperforming all baseline methods and demonstrating stronger robustness. The source code and dataset will be made publicly available.
Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
arXiv:2606.09615 [cs.RO]
(or arXiv:2606.09615v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2606.09615 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.
Full text under the publisher's open licence · 303 words
Related coverage
DexPIE: Stable Dexterous Policy Improvement from Real-World Experience
arXiv · Sep 21, 2026
SERNF: Sample-Efficient Real-World Dexterous Policy Fine-Tuning via Action-Chunked Critics and Normalizing Flows
arXiv · Jun 8, 2026
The Machines that Make the Machines
NVIDIA · Oct 7, 2026
BORA: Bridging Offline Reinforcement Learning and Online Residual Adaptation for Real-World Dexterous VLA Models
arXiv · Jun 9, 2026