Representation Learning for Dynamic Graphs: A Survey

TLDR

Graphs naturally model social networks, recommender systems, ontologies, biology, and finance, yet most machine‑learning approaches assume static structure, whereas real‑world graphs evolve over time, creating challenges for learning and inference. This survey reviews recent advances in representation learning for dynamic graphs, including dynamic knowledge graphs. The authors analyze encoder‑decoder models, categorize encoders and decoders by technique, evaluate approaches, and discuss applications, datasets, and future research directions.

Abstract

Graphs arise naturally in many real-world applications including social networks, recommender systems, ontologies, biology, and computational finance. Traditionally, machine learning models for graphs have been mostly designed for static graphs. However, many applications involve evolving graphs. This introduces important challenges for learning and inference since nodes, attributes, and edges change over time. In this survey, we review the recent advances in representation learning for dynamic graphs, including dynamic knowledge graphs. We describe existing models from an encoder-decoder perspective, categorize these encoders and decoders based on the techniques they employ, and analyze the approaches in each category. We also review several prominent applications and widely used datasets and highlight directions for future research.