Subseasonal forecasting of the weather two to six weeks in advance is critical for resource allocation and advance disaster notice but poses many challenges for the forecasting community.

We identify that the total error -- the sum of the compression errors for all iterations -- encapsulates sparsification throughout training.

A finite-state machine (FSM) is a computation model to process binary strings in sequential circuits.

Furthermore, GNNs tend topre-dominantly focus only oninformation propagation among nodes andalso belimited intheirability todescribe polyadic relationships among multiple substructures of multivariate time series or multi-node interactions in dynamics graphs.

Most efforts either focus on the high-levelfeatures associated with the last layers, orattempt tointerpret the output ofasingle layer.