Neural Message Passing on Structural Interaction Graphs for Fully-Inductive Graph Neural Networks
Omer Yom-Tov, Avigdor Gal
cs.LG
Submitted: 2026-08-13
Updated: 2026-08-14
Comments: 7 pages, 1 figure in the main body. 12 pages, 5 figures in appendix. Submitted to AAAI 2027 (main track) and currently under review
License: http://creativecommons.org/licenses/by/4.0/
The gist: A central obstacle in building graph foundation models is the input heterogeneity in terms of feature space dimensionality, semantics, and structure.
Terminology
Abstract
A central obstacle in building graph foundation models is the input heterogeneity in terms of feature space dimensionality, semantics, and structure. Such heterogeneity limits the capability of graph neural networks to generalize to new graphs with unseen feature spaces. We address the transferability challenge with SIGIL, a framework that maps any attributed graph to a unified representation space of fixed dimension. Given a graph, SIGIL lifts it to a structural interaction graph, where nodes are the input feature dimensions and weighted, typed edges encode feature alignment across multiple orders of the graph's connectivity. A relational message-passing network embeds each feature dimension into a shared space, transforming the original node features, of arbitrary dimensionality, into representations transferable to any downstream graph. By construction, SIGIL is equivariant to permutations of nodes, feature dimensions, and labels. Additionally, when the input features are one-hot indicators of discrete relations, SIGIL recovers and strictly generalizes existing foundation models for knowledge graph reasoning. A single SIGIL model, pretrained on one graph, delivers strong fully-inductive link prediction. Also, SIGIL can be used to implement existing knowledge graph foundation models. As such, SIGIL unifies several existing regimes in graph foundation model design under a single framework
Sources
- Learning Posterior Predictive Distributions for Node Classification from Synthetic Graph Priors
- Fast Graph Representation Learning with PyTorch Geometric
- Towards Foundation Models on Graphs: An Analysis on Cross-Dataset Transfer of Pretrained GNNs
- An Open Challenge for Inductive Link Prediction on Knowledge Graphs
- Open Graph Benchmark: Datasets for Machine Learning on Graphs
- TFMLinker: Universal Link Predictor by Graph In-Context Learning with Tabular Foundation Models
- Wiki-CS: A Wikipedia-Based Benchmark for Graph Neural Networks
- Pitfalls of Graph Neural Network Evaluation
- AnyGraph: Graph Foundation Model in the Wild
- A Multi-Task Perspective for Link Prediction with New Relation Types and Nodes
- Double Equivariance for Inductive Link Prediction for Both New Nodes and New Relation Types
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks