Low-rank Adaptation for Spatio-Temporal Forecasting

arXiv:2404.07919v1 Announce Type: new
Abstract: Spatio-temporal forecasting is crucial in real-world dynamic systems, predicting future changes using historical data from diverse locations. Existing methods often prioritize the development of intricate neural networks to capture the complex dependencies of the data, yet their accuracy fails to show sustained improvement. Besides, these methods also overlook node heterogeneity, hindering customized prediction modules from handling diverse regional nodes effectively. In this paper, our goal is not to propose a new model but to present …

abstract accuracy arxiv cs.ai cs.lg data dependencies development diverse dynamic forecasting future historical data improvement locations low low-rank adaptation networks neural networks node show systems temporal type world