Elucidating graph neural networks, transformers, and graph transformers

Preprint

This paper aims to present an overview of graph representation learning, delve into traditional GNNs, revisit the Transformer architecture, and explore the adaptation of Transformers for graphs. Additionally, we seek to examine the relationship between Graph Transformers and traditional GNNs. We discuss all these topics from a unified perspective, as we believe there is a lack of resources in the literature that consolidate these concepts into a single manuscript. We presume the reader is familiar with Deep Learning.

Authors: Sáez De Ocáriz Borde, H

• Publication status: Published
• Preprint server: ResearchGate
• Publication date: 2024-02-01
• DOI: 10.13140/RG.2.2.19273.31848
• Language: English
• Keywords: SBTMR