Standard direct forecasting models typically rely on point-wise objectives such as Mean Squared Error (MSE), which fail to capture the complex spatio-temporal dependencies inherent in graph-structured signals. A new study introduces FreST Loss, a frequency-enhanced spatio-temporal training objective that extends supervision to the joint spatio-temporal spectrum.

FreST Loss: A Novel Approach

FreST Loss leverages the Joint Fourier Transform (JFT) to align model predictions with ground truth in a unified spectral domain, effectively decorrelating complex dependencies across both space and time. Theoretical analysis shows that this formulation reduces estimation bias associated with time-domain training objectives.

Experimental Results

Extensive experiments on six real-world datasets demonstrate that FreST Loss is model-agnostic and consistently improves state-of-the-art baselines by better capturing holistic spatio-temporal dynamics. This approach offers a significant improvement over existing methods, addressing limitations in capturing complex spatio-temporal dependencies.