CASTANET：利用交通事件影响的因果感知时空对抗网络
CASTANET: Causality-Aware Spatio-Temporal Adversarial Network Using Traffic Incident Effects
- Sumitomo Electric System Solutions Co., Ltd.（住友电工系统解决方案有限公司）
- Kyoto University（京都大学）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
该研究针对事件引发的非周期性交通拥堵预测难题，提出融合时空图神经网络与因果处理效应估计的CASTANET，在东京数据上较基线降低RMSE，提升了拥堵预测性能。
AI中文摘要：
预测由突发事件（如事故、道路损坏）引发的非周期性交通拥堵对高级智能交通系统至关重要。然而，事件引发的拥堵难以预测，因为事件极为稀疏，发生在特定时间和地点，且影响因交通背景而异。尽管近期深度学习方法显著提升了周期性交通预测的性能，但其在非周期性拥堵预测上的表现仍有限，部分原因是未明确纳入事件记录，且事件的发生存在强烈的时空偏差。为应对这些挑战，我们提出了CASTANET，该模型融合时空图神经网络与因果处理效应估计，以利用事件记录并缓解选择偏差。在东京真实交通数据和事故记录（视为事件）上的实验表明，与最佳基线相比，CASTANET总体将RMSE降低了4.0%，在事件条件评估上降低了10.1%，在严重拥堵下的提升达到14.55%。
英文摘要：
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.
