A Generalization of Transformer Networks to Graphs

TL;DR

This paper introduces a generalized transformer architecture tailored for arbitrary graphs, incorporating neighborhood-aware attention, Laplacian eigenvector positional encoding, and edge features, improving performance on graph-based tasks.

Contribution

It presents a novel graph transformer with four key modifications, bridging the gap between traditional transformers and graph neural networks for arbitrary graph structures.

Findings

Demonstrates improved performance on graph benchmark tasks.

Shows faster training and better generalization with batch normalization.

Extends transformer capabilities to include edge feature representations.

Abstract

We propose a generalization of transformer neural network architecture for arbitrary graphs. The original transformer was designed for Natural Language Processing (NLP), which operates on fully connected graphs representing all connections between the words in a sequence. Such architecture does not leverage the graph connectivity inductive bias, and can perform poorly when the graph topology is important and has not been encoded into the node features. We introduce a graph transformer with four new properties compared to the standard model. First, the attention mechanism is a function of the neighborhood connectivity for each node in the graph. Second, the positional encoding is represented by the Laplacian eigenvectors, which naturally generalize the sinusoidal positional encodings often used in NLP. Third, the layer normalization is replaced by a batch normalization layer, which provides faster training and better generalization performance. Finally, the architecture is extended to edge feature representation, which can be critical to tasks s.a. chemistry (bond type) or link prediction (entity relationship in knowledge graphs). Numerical experiments on a graph benchmark demonstrate the performance of the proposed graph transformer architecture. This work closes the gap between the original transformer, which was designed for the limited case of line graphs, and graph neural networks, that can work with arbitrary graphs. As our architecture is simple and generic, we believe it can be used as a black box for future applications that wish to consider transformer and graphs.