Abstract:
|
Long Short--Term Memory (LSTM) is a recurrent neural network that
uses structures called memory blocks to allow the net remember
significant events distant in the past input sequence in order to
solve long time lag tasks, where other RNN approaches fail.
Throughout this work we have performed experiments using LSTM
networks extended with growing abilities, which we call GLSTM.
Four methods of training growing LSTM has been compared. These
methods include cascade and fully connected hidden layers as well
as two different levels of freezing previous weights in the
cascade case. GLSTM has been applied to a forecasting problem in a biomedical domain, where the input/output behavior of five
controllers of the Central Nervous System control has to be
modelled. We have compared growing LSTM results against other
neural networks approaches, and our work applying conventional
LSTM to the task at hand. |