Learning Posterior Predictive Distributions for Node Classification from Synthetic Graph Priors

TL;DR

NodePFN is a universal node classification method that learns from synthetic graphs to generalize across diverse real-world graphs without graph-specific training.

Contribution

Introduces NodePFN, a novel approach that trains on synthetic graphs to enable universal node classification across arbitrary graphs without additional training.

Findings

Achieves 71.27% average accuracy on 23 benchmarks.

Effectively models diverse graph properties through synthetic data.

Demonstrates strong generalization without graph-specific training.

Abstract

One of the most challenging problems in graph machine learning is generalizing across graphs with diverse properties. Graph neural networks (GNNs) face a fundamental limitation: they require separate training for each new graph, preventing universal generalization across diverse graph datasets. A critical challenge facing GNNs lies in their reliance on labeled training data for each individual graph, a requirement that hinders the capacity for universal node classification due to the heterogeneity inherent in graphs -- differences in homophily levels, community structures, and feature distributions across datasets. Inspired by the success of large language models (LLMs) that achieve in-context learning through massive-scale pre-training on diverse datasets, we introduce NodePFN. This universal node classification method generalizes to arbitrary graphs without graph-specific training. NodePFN learns posterior predictive distributions (PPDs) by training only on thousands of synthetic graphs generated from carefully designed priors. Our synthetic graph generation covers real-world graphs through the use of random networks with controllable homophily levels and structural causal models for complex feature-label relationships. We develop a dual-branch architecture combining context-query attention mechanisms with local message passing to enable graph-aware in-context learning. Extensive evaluation on 23 benchmarks demonstrates that a single pre-trained NodePFN achieves 71.27 average accuracy. These results validate that universal graph learning patterns can be effectively learned from synthetic priors, establishing a new paradigm for generalization in node classification.