|
gptkbp:instance_of
|
gptkb:neural_networks
|
|
gptkbp:architectural_style
|
hidden layers
input layer
output layer
|
|
gptkbp:characterized_by
|
gptkb:gates
cell state
forget gate
input gate
output gate
|
|
gptkbp:developed_by
|
gptkb:Jürgen_Schmidhuber
gptkb:Sepp_Hochreiter
|
|
gptkbp:first_introduced
|
gptkb:1997
|
|
gptkbp:has_limitations
|
computationally intensive
difficult to interpret
requires large datasets
|
|
gptkbp:has_variants
|
gptkb:Conv_LSTM
gptkb:Stacked_LSTM
Bidirectional LSTM
|
|
https://www.w3.org/2000/01/rdf-schema#label
|
Long Short-Term Memory Network
|
|
gptkbp:improves
|
gradient flow
|
|
gptkbp:is_a_solution_for
|
vanishing gradient problem
|
|
gptkbp:is_applied_in
|
video analysis
image captioning
music generation
|
|
gptkbp:is_compared_to
|
gptkb:Gated_Recurrent_Unit
|
|
gptkbp:is_evaluated_by
|
gptkb:historical_memory
F1 score
accuracy
precision
loss function
|
|
gptkbp:is_implemented_in
|
gptkb:Tensor_Flow
gptkb:Keras
gptkb:Py_Torch
|
|
gptkbp:is_part_of
|
gptkb:Artificial_Intelligence
gptkb:neural_networks
gptkb:machine_learning
|
|
gptkbp:is_popular_in
|
deep learning
|
|
gptkbp:is_trained_in
|
gptkb:Adam_optimizer
stochastic gradient descent
RMSprop
|
|
gptkbp:is_used_in
|
machine translation
sentiment analysis
chatbots
stock price prediction
|
|
gptkbp:requires
|
backpropagation through time
|
|
gptkbp:used_for
|
natural language processing
speech recognition
time series forecasting
sequence prediction
|
|
gptkbp:bfsParent
|
gptkb:CNTK
|
|
gptkbp:bfsLayer
|
5
|