<ul><li>Spatio-temporal forecasting is important in various fields like transportation, meteorology, and energy.</li><li>A new test-time computing paradigm called learning with calibration (ST-TTC) is introduced for spatio-temporal forecasting.</li><li>ST-TTC aims to address challenges like signal anomalies, noise, and distributional shifts by capturing periodic structural biases during testing and performing real-time bias correction.</li><li>Experiments on real-world datasets demonstrate the effectiveness, universality, flexibility, and efficiency of the ST-TTC method.</li></ul>

Learning with Calibration: Exploring Test-Time Computing of Spatio-Temporal Forecasting

Discover more