Adapting to the Stream: An Instance-Attention GNN Method for Irregular Multivariate Time Series Data
DynIMTS replaces static graphs with instance-attention that updates edge weights on the fly, delivering SOTA imputation and P12 classification ...
A team led by Guoyin Yin at Wuhan University and the Shanghai Artificial Intelligence Laboratory recently proposed a modular machine learning ...
Scholars deliver the first systematic survey of Dynamic GNNs, unifying continuous- and discrete-time models, benchmarking ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback