Détail du document
Identifiant
oai:arXiv.org:2403.11461
Sujet
Computer Science - RoboticsAuteur
Wang, Weiyao Lei, Yutian Jin, Shiyu Hager, Gregory D. Zhang, LiangjunCatégorie
Computer Science
Année
2024
Date de référencement
27/03/2024
Mots clés
in-hand tasks virtual viheRésumé
In this work, we introduce the Virtual In-Hand Eye Transformer (VIHE), a novel method designed to enhance 3D manipulation capabilities through action-aware view rendering.
VIHE autoregressively refines actions in multiple stages by conditioning on rendered views posed from action predictions in the earlier stages.
These virtual in-hand views provide a strong inductive bias for effectively recognizing the correct pose for the hand, especially for challenging high-precision tasks such as peg insertion.
On 18 manipulation tasks in RLBench simulated environments, VIHE achieves a new state-of-the-art, with a 12% absolute improvement, increasing from 65% to 77% over the existing state-of-the-art model using 100 demonstrations per task.
In real-world scenarios, VIHE can learn manipulation tasks with just a handful of demonstrations, highlighting its practical utility.
Videos and code implementation can be found at our project site: https://vihe-3d.github.io.
Wang, Weiyao,Lei, Yutian,Jin, Shiyu,Hager, Gregory D.,Zhang, Liangjun, 2024, VIHE: Virtual In-Hand Eye Transformer for 3D Robotic Manipulation
Configurable Safety Tuning of Language Models with Synthetic Preference Data
data preference cst dpo configurable language safety
Affective flexibility as a developmental building block of cognitive reappraisal: An fMRI study
cognitive success affective flexibility
