[Paper Review] A Comprehensive Survey of Graph Embedding: Problems, Techniques and Applications

Hongyun Cai, Vincent W. Zheng+1 more|arXiv (Cornell University)|Sep 22, 2017

Advanced Graph Neural Networks133 references153 citations

TL;DR

A comprehensive survey that defines graph embedding, proposes problem-setting and technique taxonomies, surveys techniques, applications, and future directions in graph embedding.

ABSTRACT

Graph is an important data representation which appears in a wide diversity of real-world scenarios. Effective graph analytics provides users a deeper understanding of what is behind the data, and thus can benefit a lot of useful applications such as node classification, node recommendation, link prediction, etc. However, most graph analytics methods suffer the high computation and space cost. Graph embedding is an effective yet efficient way to solve the graph analytics problem. It converts the graph data into a low dimensional space in which the graph structural information and graph properties are maximally preserved. In this survey, we conduct a comprehensive review of the literature in graph embedding. We first introduce the formal definition of graph embedding as well as the related concepts. After that, we propose two taxonomies of graph embedding which correspond to what challenges exist in different graph embedding problem settings and how the existing work address these challenges in their solutions. Finally, we summarize the applications that graph embedding enables and suggest four promising future research directions in terms of computation efficiency, problem settings, techniques and application scenarios.

Motivation & Objective

• Define formal graph embedding concepts and inputs/outputs.
• Propose problem-setting and technique taxonomies to organize the literature.
• Summarize applications enabled by graph embedding across node/edge/substructure/graph levels.
• Identify four promising future research directions in computation, settings, techniques, and applications.

Proposed method

• Introduce formal definitions and notations for graphs, proximity measures, and embedding objectives.
• Present two complementary taxonomies: problem settings (input/output) and embedding techniques.
• Systematically analyze how existing work addresses challenges in each setting and extract the underlying insights behind techniques.
• Categorize graph embedding applications into node-related, edge-related, and graph-related use cases with reference scenarios.
• Outline four future research directions with critical analysis of current limitations.

Experimental results

Research questions

• RQ1 What are the canonical problem settings and challenges in graph embedding across different input/output configurations?
• RQ2 What techniques address these challenges, and what insights explain why these techniques work?
• RQ3 How do graph embeddings enable various practical applications and what future directions can advance the field?

Key findings

• Two novel taxonomies organize graph embedding research by problem settings and by techniques.
• A comprehensive analysis of inputs (homogeneous/heterogeneous graphs, graphs with auxiliary information, graphs built from non-relational data) and outputs (node, edge, hybrid, whole-graph) is provided.
• The survey synthesizes techniques and extracts the underlying insights behind them, not just the methods themselves.
• Applications are categorized into node-related, edge-related, and graph-related use cases with detailed scenarios.
• Four future directions emphasize computation efficiency, problem settings, techniques, and application areas.