arXiv：cs.LG· Seyed Mohamad Moghadas, Esther Rodrigo Bonet, Bruno Cornelis, Adrian Munteanu·· 5 小时前AI 评分37
TEGER：用时空协方差实现概率交通预测的测试时校正
Teger: Spatiotemporal Covariance for Probabilistic Traffic Forecasting
AI 导读
TEGER 是一种残差协方差模型，通过闭式更新在测试时保持交通预测器的联合预测不确定性，无需重新训练。它利用固定传感器图提供低维空间精度因子，并用高斯条件化与指数移动平均波动率项校正预测均值和协方差。在四个交通数据集上，该方法改善了 60 分钟 CRPS_sum；应用于冻结的时序基础模型 Chronos 时，CRPS_sum 从 0.1798 降至 0.1736。
正文
Abstract:Traffic conditions drift -- demand patterns, incident dynamics, and sensor behavior shift over a deployment's lifetime -- so a joint uncertainty estimate fit once at training time and left static will miscalibrate as conditions change. We present TEGER, a residual covariance model that keeps a forecaster's joint predictive uncertainty current at test time through closed-form updates, not retraining. A fixed sensor graph supplies a low-dimensional spatial precision factor encoding which sensors' errors move together; at inference, Gaussian conditioning corrects The next forecast's mean and covariance from only the most recently observed residuals, and an exponential moving-average volatility term rescales marginal uncertainty to track local drift while preserving the learned correlation structure. Neither update touches the forecasting backbone's weights, so the same mechanism attaches to a frozen time-series foundation model: no gradient passes through it, turning a static point forecast into one with continually refreshed, correlated uncertainty at negligible added cost. In four traffic datasets, the fixed-graph covariance improves 60-minute $CRPS_{sum}$ over a temporal-only baseline in different configurations of the backbone and datasets. Applied post hoc to a frozen time-series foundation model, Chronos, the same test-time correction reduces $CRPS_{sum}$ from 0.1798 to 0.1736 without foundation-model fine-tuning. These results support closed-form test-time residual correction, rather than a covariance fixed once at training time.
| Comments: | Accepted in TS-LIMITS @ NeurIPS 2026 |
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2605.18068 [cs.LG] |
| (or arXiv:2605.18068v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2605.18068 arXiv-issued DOI via DataCite |
Submission history
From: Seyed Mohamad Moghadas [view email]
[v1]
Mon, 18 May 2026 08:51:33 UTC (22,652 KB)
[v2]
Thu, 1 Oct 2026 07:52:09 UTC (1,733 KB)
来源：arXiv：cs.LG · arxiv.org