====
论文
====

======
Python
=====
=

=========
Tensorflow
=========

=======
PyTorch
=======

=====
Keras
=====

====
专题
====

====
链接
====

====
视频

====

=======
药物设计

=======

=======
材料科学
=======

============
经济学与金融学
==========
==


==========
图神经网络
==========



PyG spektral DGL 图卷积网络 图注意力网络
图自编码器 图生成网络 图时空网络 图神经网路与药物设计 图网络数据
         

官网 介绍 教程 视频 FAQ
GitHub 例子     Spektral要览

Getting started

Graphs

Adjacency matrix

Node features

Edge features

Labels 

Datasets

Creating a GNN

Training the GNN
Loaders

Evaluating the GNN

Node-level learning


Data modes

Single mode Disjoint mode

Batch mode

Mixed mode


Creating a Custom Dataset

Essential information Example

Creating a Message-Passing Layer

Essential information

Example

Notes


Examples

Node-level prediction

Citation networks with GCN Citation networks with GCN (custom training loop)
Citation networks with ChebConv Citation networks with GAT
Citation networks with GAT (custom training loop) Citation networks with ARMA
Citation networks with SimpleGCN (custom transform) Open Graph Benchmark dataset

Graph-level prediction

General GNN Custom dataset
OGB mol-hiv classification (edge attributes) QM9 regression with ECC (custom training loop)
QM9 regression with ECC (batch mode) TUDataset classification with GIN
TUDataset classification with MinCut pooling  

Other applications

Graph signal classification on MNIST (mixed mode) Node clustering on citation networks with MinCut pooling (unsupervised)

Convolutional layers

MessagePassing

AGNNConv

APPNPConv

ARMAConv

ChebConv

CrystalConv

DiffusionConv

ECCConv

EdgeConv

GATConv

GatedGraphConv

GCNConv

GeneralConv

GCSConv

GINConv

GraphSageConv

TAGConv

XENetConv


Pooling layers

DiffPool

MinCutPool

SAGPool

TopKPool


Global pooling layers

GlobalAvgPool

GlobalMaxPool

GlobalSumPool

GlobalAttentionPool

GlobalAttnSumPool

SortPool


Base layers

InnerProduct

Disjoint2Batch

MinkowskiProduct


Models

GCN

GeneralGNN

GNNExplainer


Containers

Graph

Dataset

Data utils

to_disjoint

to_batch

to_mixed

batch_generator

to_tf_signature


Datasets

Citation

GraphSage

PPI

Reddit

MNIST

ModelNet

OGB

QM7

QM9

TUDataset


Loaders

Loader

SingleLoader

DisjointLoader

BatchLoader

PackedBatchLoader

MixedLoader


Transforms

AdjToSpTensor

ClusteringCoeff

Constant

Delaunay

Degree

GCNFilter

LayerPreprocess

NormalizeAdj

NormalizeOne

NormalizeSphere

OneHotLabels

 

Convolution

degree_matrix

degree_power

normalized_adjacency

laplacian

normalized_laplacian

rescale_laplacian

add_self_loops

gcn_filter

chebyshev_polynomial

chebyshev_filter


Miscellaneous

pad_jagged_array

one_hot

label_to_one_hot

flatten_list


External resources


图神经网络

什么是图神经网路,一下讲不清楚。

上海市浦东新区沪城环路999号