Neural Descriptor Fields: SE(3)-Equivariant Object Representations for Manipulation

Citations Over TimeTop 1% of 2022 papers

Abstract

We present Neural Descriptor Fields (NDFs), an object representation that encodes both points and relative poses between an object and a target (such as a robot gripper or a rack used for hanging) via category-level descriptors. We employ this representation for object manipulation, where given a task demonstration, we want to repeat the same task on a new object instance from the same category. We propose to achieve this objective by searching (via optimization) for the pose whose descriptor matches that observed in the demonstration. NDFs are conveniently trained in a self-supervised fashion via a 3D auto-encoding task that does not rely on expert-labeled keypoints. Further, NDFs are SE(3)-equivariant, guaranteeing performance that generalizes across all possible 3D object translations and rotations. We demonstrate learning of manipulation tasks from few (∼5-10) demonstrations both in simulation and on a real robot. Our performance generalizes across both object instances and 6-DoF object poses, and significantly outperforms a recent baseline that relies on 2D descriptors. Project website: https://yilundu.github.io/ndf/

Related Papers

• → Recognition by functional parts [function-based object recognition(1994)17 cited
• → Object recognition using appearance models accumulated into environment(2002)12 cited
• The Stability of Equivariant Bifurcation Problems with Parameters Symmetry and Their Unfoldings(2004)
• Unfoldings of Equivariant Bifurcation Problems with Parameter Symmetry under a Left-Right Equivalent Group(2007)
• The Unfolding of Equivariant Bifurcation Problems with Two Types of State Variables in the Presence of Parameter Symmetry(2006)