Set2Graph: Learning Graphs From Sets

Serviansky, Hadar; Segol, Nimrod; Shlomi, Jonathan; Cranmer, Kyle; Gross, Eilam; Maron, Haggai; Lipman, Yaron

Set2Graph: Learning Graphs From Sets

Feb 20, 2020

13 pages

e-Print:

2002.08772 [cs.LG]

View in:

ADS Abstract Service

pdf

reference search13 citations

Citations per year

Abstract: (submitter)

Many problems in machine learning (ML) can be cast as learning functions from sets to graphs, or more generally to hypergraphs; in short, Set2Graph functions. Examples include clustering, learning vertex and edge features on graphs, and learning triplet data in a collection. Current neural network models that approximate Set2Graph functions come from two main ML sub-fields: equivariant learning, and similarity learning. Equivariant models would be in general computationally challenging or even infeasible, while similarity learning models can be shown to have limited expressive power. In this paper we suggest a neural network model family for learning Set2Graph functions that is both practical and of maximal expressive power (universal), that is, can approximate arbitrary continuous Set2Graph functions over compact sets. Testing our models on different machine learning tasks, including an application to particle physics, we find them favorable to existing baselines.

References(49)

Figures(45)

[1]

siamese

J. Bromley
,
I. Guyon
,
Y. LeCun
,
E. Säckinger
,
R. Shah

[2]

On the equivalence between graph isomorphism testing and function approximation with gnns. arXiv preprint

Z. Chen
,
S. Villar
,
L. Chen
,
J. Bruna

•
e-Print:
- 1905.12560

[3]

Group equivariant convolutional networks. In International conference on machine learning, pages 2990-2999

T. Cohen
,
M. Welling

[4]

Spherical CNNs

e-Print:
- 1801.10130

[5]

Steerable CNNs. :1-14, 2016

T.S. Cohen
,
M. Welling

[6]

Computational geometry. In Computational geometry, pages 1-17

M. De Berg
,
M. Van Kreveld
,
M. Overmars
,
O. Schwarzkopf

[7]

Delphes 3: a modular framework for fast simulation of a generic collider experiment. Journal of High Energy Physics,(2), Feb

J. de Favereau
,
C. Delaere
,
P. Demin
,
A. Giammanco
,
V. Lemaître

et al.

[8]

Exploiting cyclic symmetry in convolutional neural networks. arXiv preprint

S. Dieleman
,
J. De Fauw
,
K. Kavukcuoglu

•
e-Print:
- 1602.02660

[9]

3d object classification and retrieval with spherical cnns. arXiv preprint

C. Esteves
,
C. Allen-Blanchette
,
A. Makadia
,
K. Daniilidis

•
e-Print:
- 1711.06721

[10]

Neural message passing for quantum chemistry. In International Conference on Machine Learning, pages 1263-1272

J. Gilmer
,
S.S. Schoenholz
,
P.F. Riley
,
O. Vinyals
,
G.E. Dahl

[11]

Attention-based deep multiple instance learning. arXiv preprint

M. Ilse
,
J.M. Tomczak
,
M. Welling

•
e-Print:
- 1802.04712

[12]

Meta-learning to cluster. arXiv preprint

Y. Jiang
,
N. Verma

•
e-Print:
- 1910.14134

[13]

Universal invariant and equivariant graph neural networks. CoRR, abs/

N. Keriven
,
G. Peyr

•
e-Print:
- 1905.04943

[14]

Adam: A Method for Stochastic Optimization

Diederik P. Kingma
(
- Amsterdam U.
)
,
Jimmy Ba
(
- Toronto U.
)

- Published as a conference paper at the 3rd International Conference for Learning Representations, San Diego, 2015
•
e-Print:
- 1412.6980

[15]

Neural relational inference for interacting systems. arXiv preprint

T. Kipf
,
E. Fetaya
,
K.-C. Wang
,
M. Welling
,
R. Zemel

•
e-Print:
- 1802.04687

[16]

Semi-Supervised Classification with Graph Convolutional Networks

e-Print:
- 1609.02907

[17]

Variational graph auto-encoders. arXiv preprint

T.N. Kipf
,
M. Welling

•
e-Print:
- 1611.07308

[18]

Covariant compositional networks for learning graphs. arXiv preprint

R. Kondor
,
H.T. Son
,
H. Pan
,
B. Anderson
,
S. Trivedi

•
e-Print:
- 1801.02144

[19]

On the generalization of equivariance and convolution in neural networks to the action of compact groups. arXiv preprint

R. Kondor
,
S. Trivedi

•
e-Print:
- 1802.03690

[20]

Imagenet classification with deep convolutional neural networks. In Advances in neural information processing systems, pages 1097-1105

A. Krizhevsky
,
I. Sutskever
,
G.E. Hinton

[21]

Gradient-based learning applied to document recognition

Y. LeCun
,
L. Bottou
,
Y. Bengio
,
P. Haffner

- IEEE Proc. 86 (1998) 2278-2324

[22]

Provably powerful graph networks. arXiv preprint

H. Maron
,
H. Ben-Hamu
,
H. Serviansky
,
Y. Lipman

•
e-Print:
- 1905.11136

[23]

Invariant and equivariant graph networks. In International Conference on Learning Representations

H. Maron
,
H. Ben-Hamu
,
N. Shamir
,
Y. Lipman

[24]

On the universality of invariant networks. arXiv preprint

H. Maron
,
E. Fetaya
,
N. Segol
,
Y. Lipman

•
e-Print:
- 1901.09342

[25]

On learning sets of symmetric elements. arXiv preprint

H. Maron
,
O. Litany
,
G. Chechik
,
E. Fetaya

•
e-Print:
- 2002.08599

1-25 of 49
1
2
25 / page