Enregistré dans:
Détails bibliographiques
Auteurs principaux: Wang, Jenny, Donca, Octavian, Held, David
Format: Preprint
Publié: 2024
Sujets:
Accès en ligne:https://arxiv.org/abs/2405.04609
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866909194962075648
author Wang, Jenny
Donca, Octavian
Held, David
author_facet Wang, Jenny
Donca, Octavian
Held, David
contents Relative placement tasks are an important category of tasks in which one object needs to be placed in a desired pose relative to another object. Previous work has shown success in learning relative placement tasks from just a small number of demonstrations when using relational reasoning networks with geometric inductive biases. However, such methods cannot flexibly represent multimodal tasks, like a mug hanging on any of n racks. We propose a method that incorporates additional properties that enable learning multimodal relative placement solutions, while retaining the provably translation-invariant and relational properties of prior work. We show that our method is able to learn precise relative placement tasks with only 10-20 multimodal demonstrations with no human annotations across a diverse set of objects within a category.
format Preprint
id arxiv_https___arxiv_org_abs_2405_04609
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Learning Distributional Demonstration Spaces for Task-Specific Cross-Pose Estimation
Wang, Jenny
Donca, Octavian
Held, David
Robotics
Relative placement tasks are an important category of tasks in which one object needs to be placed in a desired pose relative to another object. Previous work has shown success in learning relative placement tasks from just a small number of demonstrations when using relational reasoning networks with geometric inductive biases. However, such methods cannot flexibly represent multimodal tasks, like a mug hanging on any of n racks. We propose a method that incorporates additional properties that enable learning multimodal relative placement solutions, while retaining the provably translation-invariant and relational properties of prior work. We show that our method is able to learn precise relative placement tasks with only 10-20 multimodal demonstrations with no human annotations across a diverse set of objects within a category.
title Learning Distributional Demonstration Spaces for Task-Specific Cross-Pose Estimation
topic Robotics
url https://arxiv.org/abs/2405.04609