Graph-aware networks improve zero-shot traffic forecasting accuracy
GAS-PFN: Graph-Aware Spatiotemporal Prior-Data Fitted Networks for Zero-Shot Traffic Forecasting
Machine LearningArtificial IntelligenceNeural and Evolutionary Computing
Summary
Traffic forecasting is tough because it depends on how roads connect and how traffic changes over time. The authors present a new approach called GAS-PFN that teaches a model to understand both road layouts and traffic patterns without needing retraining for each new city. Their method uses a special way of showing the model how sensors on roads interact over time, improving its predictions. They tested it on six different scenarios and found it predicted traffic more accurately than earlier zero-shot methods.
What this means in practice
- •For transportation planners: Predict traffic in new or changing road networks without retraining models for each location using graph-aware zero-shot forecasting.
- •For mobile network operators: Improve wireless network traffic forecasts by modeling spatiotemporal dynamics over sensor-like network measurement points with zero-shot adaptability.
Authors
Joseph Natter, Long Vu, Yifan Zhang, Bobin Deng, Kun Suo, Rui Wu
Abstract
Traffic forecasting requires modeling complex temporal dynamics and the spatial dependencies imposed by road topology. Supervised graph-based models achieve strong performance, but are often trained for a particular dataset and sensor graph, making deployment to new or evolving networks costly. Prior-Data Fitted Networks (PFNs) offer a zero-shot alternative by learning in-context inference from synthetic task priors, which provide an explicit mechanism for encoding domain-specific assumptions. Existing forecasting PFNs, however, do not explicitly represent the graph-structured dependencies governing traffic dynamics. To combine graph-aware traffic modeling with zero-shot adaptation, we introduce the Graph-Aware Spatiotemporal Prior-Data Fitted Network (GAS-PFN). GAS-PFN incorporates sampled road topology into its synthetic task prior through random-walk diffusion, teaching the model to learn graph-coupled dynamics during pretraining. Its co-designed inference architecture uses two mask-conditioned passes through a shared Transformer backbone to separately model sensor-specific temporal dynamics and spatiotemporal interactions among neighboring sensors before integrating their representations for prediction. This design enables GAS-PFN to use target-network structure and context without parameter updates. Across six benchmarks at the 90-minute horizon, GAS-PFN reduces MSE by 34.0% on average relative to TimePFN, its closest PFN-based predecessor. It also reduces MSE by 27.7% on average relative to four recent industry-developed general-purpose zero-shot forecasters. In this comparison, all models receive 90 minutes of contiguous recent history, and any additional context is drawn from each dataset's chronological training split. These results indicate that aligning PFN priors and inference architectures with graph structure improves in-context learning for zero-shot traffic forecasting.