Temporal Graph MLP Mixer for Spatio-Temporal Forecasting
This work addresses robust forecasting for applications like traffic prediction and climate modeling, but it is incremental as it builds on existing methods without achieving broad SOTA.
The paper tackles spatiotemporal forecasting with missing data by introducing the Temporal Graph MLP-Mixer (T-GMM), which combines node-level and patch-level processing to capture dependencies, and demonstrates effective forecasting on datasets like AQI and METR-LA, though it does not surpass state-of-the-art in all scenarios.
Spatiotemporal forecasting is critical in applications such as traffic prediction, climate modeling, and environmental monitoring. However, the prevalence of missing data in real-world sensor networks significantly complicates this task. In this paper, we introduce the Temporal Graph MLP-Mixer (T-GMM), a novel architecture designed to address these challenges. The model combines node-level processing with patch-level subgraph encoding to capture localized spatial dependencies while leveraging a three-dimensional MLP-Mixer to handle temporal, spatial, and feature-based dependencies. Experiments on the AQI, ENGRAD, PV-US and METR-LA datasets demonstrate the model's ability to effectively forecast even in the presence of significant missing data. While not surpassing state-of-the-art models in all scenarios, the T-GMM exhibits strong learning capabilities, particularly in capturing long-range dependencies. These results highlight its potential for robust, scalable spatiotemporal forecasting.