Graph Transformer

Introduced by Dwivedi et al. in A Generalization of Transformer Networks to Graphs

This is Graph Transformer method, proposed as a generalization of Transformer Neural Network architectures, for arbitrary graphs.

Compared to the original Transformer, the highlights of the presented architecture are:

The attention mechanism is a function of neighborhood connectivity for each node in the graph.
The position encoding is represented by Laplacian eigenvectors, which naturally generalize the sinusoidal positional encodings often used in NLP.
The layer normalization is replaced by a batch normalization layer.
The architecture is extended to have edge representation, which can be critical to tasks with rich information on the edges, or pairwise interactions (such as bond types in molecules, or relationship type in KGs. etc).

Source: A Generalization of Transformer Networks to Graphs

Read Paper See Code

Paper	Code	Results	Date	Stars

Task	Papers	Share
Node Classification	18	5.88%
Graph Representation Learning	16	5.23%
Graph Learning	14	4.58%
Graph Classification	12	3.92%
Link Prediction	11	3.59%
Graph Regression	11	3.59%
Property Prediction	8	2.61%
Drug Discovery	8	2.61%
Molecular Property Prediction	7	2.29%

This feature is experimental; we are continuously improving our matching algorithm.

Component	Type	Add Remove
Laplacian PE	Graph Embeddings
Transformer	Transformers