ST-LoRA: Low-rank Adaptation for Spatio-Temporal Forecasting
This work addresses inefficiencies in spatio-temporal forecasting models for applications like urban planning or environmental monitoring, offering a computationally efficient solution with incremental improvements.
The paper tackles the problem of node-level heterogeneity and over-parameterization in spatio-temporal forecasting by introducing a low-rank adaptation framework that adjusts existing models with node-level components. The result is a method that improves performance by 7% with only 1% additional parameter cost across multiple real-world datasets.
Spatio-temporal forecasting is essential for understanding future dynamics within real-world systems by leveraging historical data from multiple locations. Existing methods often prioritize the development of intricate neural networks to capture the complex dependencies of the data. These methods neglect node-level heterogeneity and face over-parameterization when attempting to model node-specific characteristics. In this paper, we present a novel low-rank adaptation framework for existing spatio-temporal prediction models, termed \model, which alleviates the aforementioned problems through node-level adjustments. Specifically, we introduce the node-adaptive low-rank layer and node-specific predictor, capturing the complex functional characteristics of nodes while maintaining computational efficiency. Extensive experiments on multiple real-world datasets demonstrate that our method consistently achieves superior performance across various forecasting models with minimal computational overhead, improving performance by 7% with only 1% additional parameter cost. The source code is available at https://github.com/RWLinno/ST-LoRA.