aiqbench analysis
CASTANET applies spatio-temporal graph neural networks and causal inference to improve traffic congestion forecasting specifically during sudden incidents like accidents. By explicitly modeling the selection bias inherent in sparse incident data, this approach offers a more robust alternative to standard deep learning models for non-periodic traffic patterns. The method demonstrates significant performance gains on real-world Tokyo datasets, highlighting its potential for advanced intelligent transportation systems.
CASTANET: Causality-Aware Spatio-Temporal Adversarial Network Using Traffic Incident Effects
arXiv:2608.27942v1 Announce Type: new Abstract: Predicting non-periodic traffic congestion caused by sudden incidents (e.g., accidents and road damage) is crucial for advanced intelligent transportation systems. However, incident-driven congestion is difficult to forecast because incidents are extremely sparse, occur at specific times and locations, and have heterogeneous impacts depending on the traffic context. While recent deep learning approaches have significantly improved periodic traffic forecasting, their performance on non-periodic congestion remains limited, partly because incident records are not explicitly incorporated and their occurrence is strongly biased in space and time. To address these challenges, we propose CASTANET, which integrates spatio-temporal graph neural networks and causal treatment effect estimation to utilize incident records while mitigating selection bias. Experiments on real-world traffic data and accident records from Tokyo, which we treat as incidents, show that CASTANET reduces RMSE by 4.0% overall compared to the best baseline and by 10.1% on incident-conditioned evaluation, with gains reaching 14.55% under severe congestion.
Abstract:Predicting non-periodic traffic congestion caused by sudden incidents (e.g., accidents and road damage) is crucial for advanced intelligent transportation systems. However, incident-driven congestion is difficult to forecast because incidents are extremely sparse, occur at specific times and locations, and have heterogeneous impacts depending on the traffic context. While recent deep learning approaches have significantly improved periodic traffic forecasting, their performance on non-periodic congestion remains limited, partly because incident records are not explicitly incorporated and their occurrence is strongly biased in space and time. To address these challenges, we propose CASTANET, which integrates spatio-temporal graph neural networks and causal treatment effect estimation to utilize incident records while mitigating selection bias. Experiments on real-world traffic data and accident records from Tokyo, which we treat as incidents, show that CASTANET reduces RMSE by 4.0% overall compared to the best baseline and by 10.1% on incident-conditioned evaluation, with gains reaching 14.55% under severe congestion.
Artificial Intelligence (cs.AI)
arXiv:2608.27942 [cs.AI]
(or arXiv:2608.27942v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2608.27942 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 244 words
Related coverage
Hybrid Robustness Verification for Spatio-Temporal Neural Networks
arXiv · Jun 9, 2026
Dynamic Link Prediction with Temporally Enhanced Signed Graph Neural Networks
arXiv · Jun 16, 2026
Spatio-Temporal Fusion Model for Standard View Classification of Echocardiographic Videos
arXiv · Jun 17, 2026
Spatio-Temporal Attention Graph Neural Network: Explaining Causalities With Attention
arXiv · Jun 10, 2026