-
ALSTM contains a temporal attentive aggregation layer based on normal LSTM.
-
Paper: A dual-stage attention-based recurrent neural network for time series prediction.
-
NOTE: Current version of implementation is just a simplified version of ALSTM. It is an LSTM with attention.