benedekrozemberczki/GraphWaveletNeuralNetwork benedekrozemberczki/GraphWaveletNeuralNetwork

  • Stars
    star
    548
  • Rank 78,295 (Top 2 %)
  • Language
    Python
  • License
    GNU General Publi...
  • Created over 5 years ago
  • Updated about 1 year ago
Twitter logo Share LinkedIn logo Share Facebook logo Share
benedekrozemberczki/GraphWaveletNeuralNetwork
github

benedekrozemberczki

Reviews

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

Repository Details

A PyTorch implementation of "Graph Wavelet Neural Network" (ICLR 2019)

Graph Wavelet Neural Network

Arxiv codebeat badge repo size benedekrozemberczki

A PyTorch implementation of Graph Wavelet Neural Network (ICLR 2019).

Abstract

We present graph wavelet neural network (GWNN), a novel graph convolutional neural network (CNN), leveraging graph wavelet transform to address the shortcomings of previous spectral graph CNN methods that depend on graph Fourier transform. Different from graph Fourier transform, graph wavelet transform can be obtained via a fast algorithm without requiring matrix eigendecomposition with high computational cost. Moreover, graph wavelets are sparse and localized in vertex domain, offering high efficiency and good interpretability for graph convolution. The proposed GWNN significantly outperforms previous spectral graph CNNs in the task of graph-based semi-supervised classification on three benchmark datasets: Cora, Citeseer and Pubmed.

A reference Tensorflow implementation is accessible [here].

This repository provides an implementation of Graph Wavelet Neural Network as described in the paper:

Graph Wavelet Neural Network. Bingbing Xu, Huawei Shen, Qi Cao, Yunqi Qiu, Xueqi Cheng. ICLR, 2019. [Paper]

Requirements

The codebase is implemented in Python 3.5.2. package versions used for development are just below.

networkx          2.4
tqdm              4.28.1
numpy             1.15.4
pandas            0.23.4
texttable         1.5.0
scipy             1.1.0
argparse          1.1.0
torch             1.1.0
torch-scatter     1.4.0
torch-sparse      0.4.3
torch-cluster     1.4.5
torch-geometric   1.3.2
torchvision       0.3.0
scikit-learn      0.20.0
PyGSP             0.5.1

Datasets

The code takes the **edge list** of the graph in a csv file. Every row indicates an edge between two nodes separated by a comma. The first row is a header. Nodes should be indexed starting with 0. A sample graph for `Cora` is included in the `input/` directory. In addition to the edgelist there is a JSON file with the sparse features and a csv with the target variable.

The **feature matrix** is a sparse binary one it is stored as a json. Nodes are keys of the json and feature indices are the values. For each node feature column ids are stored as elements of a list. The feature matrix is structured as:

{ 0: [0, 1, 38, 1968, 2000, 52727],
  1: [10000, 20, 3],
  2: [],
  ...
  n: [2018, 10000]}

The **target vector** is a csv with two columns and headers, the first contains the node identifiers the second the targets. This csv is sorted by node identifiers and the target column contains the class meberships indexed from zero.

NODE ID Target
0 3
1 1
2 0
3 1
... ...
n 3

Options

Training the model is handled by the src/main.py script which provides the following command line arguments.

Input and output options

  --edge-path        STR   Input graph path.   Default is `input/cora_edges.csv`.
  --features-path    STR   Features path.      Default is `input/cora_features.json`.
  --target-path      STR   Target path.        Default is `input/cora_target.csv`.
  --log-path         STR   Log path.           Default is `logs/cora_logs.json`.

Model options

  --epochs                INT       Number of Adam epochs.         Default is 200.
  --learning-rate         FLOAT     Number of training epochs.     Default is 0.01.
  --weight-decay          FLOAT     Weight decay.                  Default is 5*10**-4.
  --filters               INT       Number of filters.             Default is 16.
  --dropout               FLOAT     Dropout probability.           Default is 0.5.
  --test-size             FLOAT     Test set ratio.                Default is 0.2.
  --seed                  INT       Random seeds.                  Default is 42.
  --approximation-order   INT       Chebyshev polynomial order.    Default is 3.
  --tolerance             FLOAT     Wavelet coefficient limit.     Default is 10**-4.
  --scale                 FLOAT     Heat kernel scale.             Default is 1.0.

Examples

The following commands learn the weights of a graph wavelet neural network and saves the logs. The first example trains a graph wavelet neural network on the default dataset with standard hyperparameter settings. Saving the logs at the default path. 

python src/main.py

Training a model with more filters in the first layer.

python src/main.py --filters 32

Approximationg the wavelets with polynomials that have an order of 5.

python src/main.py --approximation-order 5

License

More Repositories

1

awesome-graph-classification

A collection of important graph embedding, classification and representation learning papers with implementations.
Python
4,666
star
2

pytorch_geometric_temporal

PyTorch Geometric Temporal: Spatiotemporal Signal Processing with Neural Machine Learning Models (CIKM 2021)
Python
2,455
star
3

awesome-decision-tree-papers

A collection of research papers on decision, classification and regression trees with implementations.
Python
2,248
star
4

awesome-community-detection

A curated list of community detection research papers with implementations.
Python
2,224
star
5

karateclub

Karate Club: An API Oriented Open-source Python Framework for Unsupervised Learning on Graphs (CIKM 2020)
Python
2,065
star
6

awesome-fraud-detection-papers

A curated list of data mining papers about fraud detection.
Python
1,481
star
7

CapsGNN

A PyTorch implementation of "Capsule Graph Neural Network" (ICLR 2019).
Python
1,216
star
8

awesome-gradient-boosting-papers

A curated list of gradient boosting research papers with implementations.
Python
966
star
9

graph2vec

A parallel implementation of "graph2vec: Learning Distributed Representations of Graphs" (MLGWorkshop 2017).
Python
860
star
10

ClusterGCN

A PyTorch implementation of "Cluster-GCN: An Efficient Algorithm for Training Deep and Large Graph Convolutional Networks" (KDD 2019).
Python
757
star
11

littleballoffur

Little Ball of Fur - A graph sampling extension library for NetworKit and NetworkX (CIKM 2020)
Python
676
star
12

SimGNN

A PyTorch implementation of "SimGNN: A Neural Network Approach to Fast Graph Similarity Computation" (WSDM 2019).
Python
657
star
13

awesome-monte-carlo-tree-search-papers

A curated list of Monte Carlo tree search papers with implementations.
Python
565
star
14

datasets

A repository of pretty cool datasets that I collected for network science and machine learning research.
551
star
15

MixHop-and-N-GCN

An implementation of "MixHop: Higher-Order Graph Convolutional Architectures via Sparsified Neighborhood Mixing" (ICML 2019).
Python
395
star
16

APPNP

A PyTorch implementation of "Predict then Propagate: Graph Neural Networks meet Personalized PageRank" (ICLR 2019).
Python
351
star
17

AttentionWalk

A PyTorch Implementation of "Watch Your Step: Learning Node Embeddings via Graph Attention" (NeurIPS 2018).
Python
309
star
18

SGCN

A PyTorch implementation of "Signed Graph Convolutional Network" (ICDM 2018).
Python
262
star
19

GAM

A PyTorch implementation of "Graph Classification Using Structural Attention" (KDD 2018).
Python
261
star
20

GEMSEC

The TensorFlow reference implementation of 'GEMSEC: Graph Embedding with Self Clustering' (ASONAM 2019).
Python
244
star
21

SEAL-CI

A PyTorch implementation of "Semi-Supervised Graph Classification: A Hierarchical Graph Perspective" (WWW 2019)
Python
204
star
22

shapley

The official implementation of "The Shapley Value of Classifiers in Ensemble Games" (CIKM 2021).
Python
203
star
23

Splitter

A Pytorch implementation of "Splitter: Learning Node Representations that Capture Multiple Social Contexts" (WWW 2019).
Python
203
star
24

DANMF

A sparsity aware implementation of "Deep Autoencoder-like Nonnegative Matrix Factorization for Community Detection" (CIKM 2018).
Python
194
star
25

GraphWaveMachine

A scalable implementation of "Learning Structural Node Embeddings Via Diffusion Wavelets (KDD 2018)".
Python
176
star
26

role2vec

A scalable Gensim implementation of "Learning Role-based Graph Embeddings" (IJCAI 2018).
Python
158
star
27

MUSAE

The reference implementation of "Multi-scale Attributed Node Embedding". (Journal of Complex Networks 2021)
Python
136
star
28

EdMot

An implementation of "EdMot: An Edge Enhancement Approach for Motif-aware Community Detection" (KDD 2019)
Python
128
star
29

M-NMF

An implementation of "Community Preserving Network Embedding" (AAAI 2017)
Python
119
star
30

diff2vec

Reference implementation of Diffusion2Vec (Complenet 2018) built on Gensim and NetworkX.
Python
117
star
31

LabelPropagation

A NetworkX implementation of Label Propagation from a "Near Linear Time Algorithm to Detect Community Structures in Large-Scale Networks" (Physical Review E 2008).
Python
111
star
32

walklets

A lightweight implementation of Walklets from "Don't Walk Skip! Online Learning of Multi-scale Network Embeddings" (ASONAM 2017).
Python
98
star
33

tigerlily

TigerLily: Finding drug interactions in silico with the Graph.
Jupyter Notebook
95
star
34

BANE

A sparsity aware implementation of "Binarized Attributed Network Embedding" (ICDM 2018).
Python
85
star
35

EgoSplitting

A NetworkX implementation of "Ego-splitting Framework: from Non-Overlapping to Overlapping Clusters" (KDD 2017).
Python
80
star
36

ASNE

A sparsity aware and memory efficient implementation of "Attributed Social Network Embedding" (TKDE 2018).
Python
77
star
37

TENE

A sparsity aware implementation of "Enhanced Network Embedding with Text Information" (ICPR 2018).
Python
71
star
38

SINE

A PyTorch Implementation of "SINE: Scalable Incomplete Network Embedding" (ICDM 2018).
Python
69
star
39

RolX

An alternative implementation of Recursive Feature and Role Extraction (KDD11 & KDD12)
Python
58
star
40

GraRep

A SciPy implementation of "GraRep: Learning Graph Representations with Global Structural Information" (WWW 2015).
Python
58
star
41

PDN

The official PyTorch implementation of "Pathfinder Discovery Networks for Neural Message Passing" (WebConf '21)
Python
55
star
42

TADW

An implementation of "Network Representation Learning with Rich Text Information" (IJCAI '15).
Python
54
star
43

spatiotemporal_datasets

Spatiotemporal datasets collected for network science, deep learning and general machine learning research.
43
star
44

NMFADMM

A sparsity aware implementation of "Alternating Direction Method of Multipliers for Non-Negative Matrix Factorization with the Beta-Divergence" (ICASSP 2014).
Python
40
star
45

FEATHER

The reference implementation of FEATHER from the CIKM '20 paper "Characteristic Functions on Graphs: Birds of a Feather, from Statistical Descriptors to Parametric Models".
Python
40
star
46

BoostedFactorization

An implementation of "Multi-Level Network Embedding with Boosted Low-Rank Matrix Approximation" (ASONAM 2019).
Python
33
star
47

resolutions-2019

A list of data mining and machine learning papers that I implemented in 2019.
20
star
48

OrbitalFeatures

A sparsity aware implementation of "Biological Network Comparison Using Graphlet Degree Distribution" (Bioinformatics 2007)
Python
19
star
49

FSCNMF

An implementation of "Fusing Structure and Content via Non-negative Matrix Factorization for Embedding Information Networks".
Python
18
star
50

GRAF

Inner product natural graph factorization machine used in 'GEMSEC: Graph Embedding with Self Clustering' .
Python
10
star
51

HullCoverConditionedUnitDiskGraph

A generator for unit disk graphs conditioned on concave hull cover.
Python
8
star
52

AV_Ultimate_Student_Hunt

Solution for the Ultimate Student Hunt Challenge (1st place).
R
8
star
53

NestedSubtreeHash

A distributed implementation of "Nested Subtree Hash Kernels for Large-Scale Graph Classification Over Streams" (ICDM 2012).
Python
7
star
54

Societe-General

Solution for ENS - Societe Generale Challenge (1st place).
R
5
star
55

resolutions-2020

4
star
56

graphmining.ai

Benedek Rozemberczki Personal Webpage
4
star
57

benedekrozemberczki

3
star