Automated Graph Machine Learning: Approaches, Libraries, Benchmarks and Directions

TL;DR

This work surveys automated graph machine learning, focusing on hyper-parameter optimization (HPO) and neural architecture search (NAS) for graph tasks. It introduces AutoGL, the first open-source library dedicated to automated graph learning, and NAS-Bench-Graph, a benchmark for fair, reproducible graph NAS evaluations. The paper details graph-specific AutoML formulations, search spaces, strategies, and estimation techniques, highlighting recent advances and practical library support. It concludes with future directions spanning scalability, explainability, robustness, and hardware-aware design to advance automated graph learning in both academia and industry.

Abstract

Graph machine learning has been extensively studied in both academic and industry. However, as the literature on graph learning booms with a vast number of emerging methods and techniques, it becomes increasingly difficult to manually design the optimal machine learning algorithm for different graph-related tasks. To tackle the challenge, automated graph machine learning, which aims at discovering the best hyper-parameter and neural architecture configuration for different graph tasks/data without manual design, is gaining an increasing number of attentions from the research community. In this paper, we extensively discuss automated graph machine learning approaches, covering hyper-parameter optimization (HPO) and neural architecture search (NAS) for graph machine learning. We briefly overview existing libraries designed for either graph machine learning or automated machine learning respectively, and further in depth introduce AutoGL, our dedicated and the world's first open-source library for automated graph machine learning. Also, we describe a tailored benchmark that supports unified, reproducible, and efficient evaluations. Last but not least, we share our insights on future research directions for automated graph machine learning. This paper is the first systematic and comprehensive discussion of approaches, libraries as well as directions for automated graph machine learning.

Figures (4)

• Figure 1: The overall framework of AutoGL.
• Figure 2: Nine different choices of our macro search space. Each node is a representation of vertices and each edge is an operation qin2022bench.
• Figure 3: An example architecture.
• Figure 4: The learning curve depicting the optimal performance as a function of the number of searched architectures is presented herein. The reported results are obtained by averaging measurements from five independent experiments, each conducted with distinct random seeds. The background of the figure displays the standard errors associated with the reported values.