A Survey of Few-Shot Learning on Graphs: from Meta-Learning to Pre-Training and Prompt Learning

TL;DR

This survey addresses the challenge of data scarcity in graph learning by introducing two comprehensive taxonomies: problem settings (label scarcity vs structure scarcity) and technique families (meta-learning, pre-training, and hybrid approaches). It surveys core methods within each category, detailing how graph structure and prompts can enhance priors and downstream adaptation, and discusses the trade-offs between fully supervised meta-learning and self-supervised pre-training. The work highlights two parallel paths—meta-learning that relies on annotated base tasks and pre-training that leverages unlabeled data, supplemented by hybrid approaches that combine both. It also identifies open challenges, such as scaling to large and complex graphs, cross-domain transfer, and interpretable prompting, and outlines directions toward graph foundation models and broader task coverage. Overall, the paper provides a structured, in-depth roadmap for advancing few-shot learning on graphs through rigorous taxonomy and critical synthesis.

Abstract

Graph representation learning, a critical step in graph-centric tasks, has seen significant advancements. Earlier techniques often operate in an end-to-end setting, which heavily rely on the availability of ample labeled data. This constraint has spurred the emergence of few-shot learning on graphs, where only a few labels are available for each task. Given the extensive literature in this field, this survey endeavors to synthesize recent developments, provide comparative insights, and identify future directions. We systematically categorize existing studies based on two major taxonomies: (1) Problem taxonomy, which explores different types of data scarcity problems and their applications, and (2) Technique taxonomy, which details key strategies for addressing these data-scarce few-shot problems. The techniques can be broadly categorized into meta-learning, pre-training, and hybrid approaches, with a finer-grained classification in each category to aid readers in their method selection process. Within each category, we analyze the relationships among these methods and compare their strengths and limitations. Finally, we outline prospective directions for few-shot learning on graphs to catalyze continued innovation in this field. The website for this survey can be accessed by \url{https://github.com/smufang/fewshotgraph}.

Figures (7)

• Figure 1: Overview of few-shot learning on graphs. (a--b) Different problem settings involving label or structure scarcity on graphs. (c--e) Different applications, such as cold-start recommendation on user-item graphs.
• Figure 2: Overview of few-shot learning techniques on graphs.
• Figure 3: Taxonomy of few-shot learning problems on graphs.
• Figure 4: Taxonomy of few-shot learning techniques on graphs.
• Figure 5: An illustration of meta-learning on graph.