<ul><li>Source-Free Domain Adaptation (SFDA) aims to adapt pre-trained models without accessing source data to preserve data privacy.</li><li>Existing SFDA methods struggle with multivariate time series (MTS) due to neglecting spatial correlations inherent in MTS data.</li><li>Temporal Restoration and Spatial Rewiring (TERSE) is proposed as a concise SFDA method tailored for MTS data, considering spatial correlations.</li><li>TERSE consists of a spatial-temporal feature encoder, temporal restoration, and spatial rewiring tasks to transfer dependencies across domains effectively.</li></ul>

Temporal Restoration and Spatial Rewiring for Source-Free Multivariate Time Series Domain Adaptation

Discover more