Abstract: Deep learning models have significantly addressed the challenges of multivariate time series forecasting. Recently, Transformer-based models which have primarily focused on either temporal ...