multivariate time series forecasting is extensively studied throughout the
years with ubiquitous applications in areas such as finance, traffic,
environment, etc. Still, concerns have been raised on traditional methods for
incapable of modeling complex patterns or dependencies lying in