GUIDED Network-Agnostic Feature Initialization for Spatial Transferability in GNN-based Models

2026-07-21

AI Summary

This work addresses the limited cross-city transferability of graph neural networks in traffic assignment due to their reliance on specific road network topologies. To overcome this, the authors propose GUIDED, a topology-agnostic feature initialization layer that injects origin–destination (OD) demand as scalar attributes on virtual edges, thereby standardizing the input space. This approach decouples OD demand from physical nodes for the first time, constructing structure-invariant feature embeddings via virtual links. The resulting model exhibits substantially improved out-of-distribution robustness and parameter efficiency, especially under data-scarce conditions. Integrated with a heterogeneous graph attention network (HetGAT) and an optimized scatter operation, the framework maintains state-of-the-art prediction accuracy on individual networks while significantly outperforming baselines in cross-network transfer tasks and achieving approximately 50% faster training. This paradigm offers a generalizable solution for OD-based problems such as logistics and multimodal transportation.

Abstract

The Traffic Assignment Problem is a fundamental but computationally expensive component of transportation planning. While Graph Neural Networks have emerged as fast, data-driven surrogates, their practical deployment is severely constrained by a spatial generalization gap. Standard models rely on transductive feature initializations that tie travel demand to fixed network topologies, preventing seamless transfer to new urban environments. To overcome this structural limitation, this research proposes a network-agnostic initialization layer, termed Geometrically Unconstrained Inductive Demand EmbeDding (GUIDED). By injecting travel demand as a scalar attribute on auxiliary virtual links rather than as specific node features, this modular framework standardizes the input space regardless of network scale. Extensive experimental evaluation across multiple urban topologies demonstrates that a Heterogeneous Graph Attention Network (HetGAT) model integrated with the proposed GUIDED layer maintains state-of-the-art predictive accuracy on single-network tasks, while demonstrating superior robustness to out-of-distribution demand patterns and maintaining a distinct performance advantage over the baseline even under severe data scarcity. Notably, the proposed feature initialization enables highly parameter-efficient domain adaptation for inter-network transfer learning without artificial input homogenization, establishing a robust foundation for truly inductive models. At the same time, the optimized scatter operations of the initialization layer yield an approximate 50% reduction in training time per epoch compared to the baseline approach. Furthermore, while demonstrated on vehicular traffic, this fundamental abstraction of spatial topology provides a versatile blueprint for generalized origin-destination spatial problems, such as freight logistics and multimodal network optimization.

Problem

• spatial transferability
• traffic assignment
• network-agnostic
• feature initialization
• graph neural networks

Innovation

• network-agnostic initialization
• spatial transferability
• inductive GNN
• virtual link embedding
• traffic assignment