A SIMPLE AND UNIVERSAL ROTATION EQUIVARIANT POINT-CLOUD NETWORK

Ben Finkelshtein; Chaim Baskin; Haggai Maron; Nadav Dym

A SIMPLE AND UNIVERSAL ROTATION EQUIVARIANT POINT-CLOUD NETWORK

Ben Finkelshtein, Chaim Baskin, Haggai Maron, Nadav Dym

Research output: Contribution to journal › Conference article › peer-review

5 Scopus citations

Abstract

Equivariance to permutations and rigid motions is an important inductive bias for various 3D learning problems. Recently it has been shown that the equivariant Tensor Field Network architecture is universal- it can approximate any equivariant function. In this paper we suggest a much simpler architecture, prove that it enjoys the same universality guarantees and evaluate its performance on Modelnet40. The code to reproduce our experiments is available at https://github.com/simpleinvariance/UniversalNetwork.

Original language	English
Pages (from-to)	107-115
Number of pages	9
Journal	Proceedings of Machine Learning Research
Volume	196
State	Published - 1 Jan 2022
Externally published	Yes
Event	ICML Workshop on Topology, Algebra, and Geometry in Machine Learning, TAG:ML 2022 - Virtual, Online, United States Duration: 20 Jul 2022 → …

ASJC Scopus subject areas

Artificial Intelligence
Software
Control and Systems Engineering
Statistics and Probability

Cite this

@article{7c60c3445a7b413a8ed3e0a1f4510657,

title = "A SIMPLE AND UNIVERSAL ROTATION EQUIVARIANT POINT-CLOUD NETWORK",

abstract = "Equivariance to permutations and rigid motions is an important inductive bias for various 3D learning problems. Recently it has been shown that the equivariant Tensor Field Network architecture is universal- it can approximate any equivariant function. In this paper we suggest a much simpler architecture, prove that it enjoys the same universality guarantees and evaluate its performance on Modelnet40. The code to reproduce our experiments is available at https://github.com/simpleinvariance/UniversalNetwork.",

author = "Ben Finkelshtein and Chaim Baskin and Haggai Maron and Nadav Dym",

note = "Publisher Copyright: {\textcopyright} 2022 Proceedings of Machine Learning Research. All rights reserved.; ICML Workshop on Topology, Algebra, and Geometry in Machine Learning, TAG:ML 2022 ; Conference date: 20-07-2022",

year = "2022",

month = jan,

day = "1",

language = "English",

volume = "196",

pages = "107--115",

journal = "Proceedings of Machine Learning Research",

issn = "2640-3498",

}

TY - JOUR

T1 - A SIMPLE AND UNIVERSAL ROTATION EQUIVARIANT POINT-CLOUD NETWORK

AU - Finkelshtein, Ben

AU - Baskin, Chaim

AU - Maron, Haggai

AU - Dym, Nadav

PY - 2022/1/1

Y1 - 2022/1/1

N2 - Equivariance to permutations and rigid motions is an important inductive bias for various 3D learning problems. Recently it has been shown that the equivariant Tensor Field Network architecture is universal- it can approximate any equivariant function. In this paper we suggest a much simpler architecture, prove that it enjoys the same universality guarantees and evaluate its performance on Modelnet40. The code to reproduce our experiments is available at https://github.com/simpleinvariance/UniversalNetwork.

AB - Equivariance to permutations and rigid motions is an important inductive bias for various 3D learning problems. Recently it has been shown that the equivariant Tensor Field Network architecture is universal- it can approximate any equivariant function. In this paper we suggest a much simpler architecture, prove that it enjoys the same universality guarantees and evaluate its performance on Modelnet40. The code to reproduce our experiments is available at https://github.com/simpleinvariance/UniversalNetwork.

UR - http://www.scopus.com/inward/record.url?scp=85163669450&partnerID=8YFLogxK

M3 - Conference article

AN - SCOPUS:85163669450

SN - 2640-3498

VL - 196

SP - 107

EP - 115

JO - Proceedings of Machine Learning Research

JF - Proceedings of Machine Learning Research

T2 - ICML Workshop on Topology, Algebra, and Geometry in Machine Learning, TAG:ML 2022

Y2 - 20 July 2022

ER -

A SIMPLE AND UNIVERSAL ROTATION EQUIVARIANT POINT-CLOUD NETWORK

Abstract

ASJC Scopus subject areas

Other files and links

Fingerprint

Cite this