Statements (22)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
Recurrent neural network architecture
|
| gptkbp:advantage |
Improved learning of precise timing
|
| gptkbp:allows |
Gates to access cell state directly
|
| gptkbp:basedOn |
Long Short-Term Memory
|
| gptkbp:developedBy |
gptkb:Jürgen_Schmidhuber
Felix Gers Fred Cummins |
| gptkbp:hasComponent |
Cell state
Forget gate Input gate Output gate |
| gptkbp:hasFeature |
Peephole connections
|
| https://www.w3.org/2000/01/rdf-schema#label |
Peephole LSTM
|
| gptkbp:introducedIn |
2000
|
| gptkbp:publishedIn |
Neural Computation journal
|
| gptkbp:relatedTo |
gptkb:GRU
gptkb:LSTM |
| gptkbp:usedIn |
Speech recognition
Sequence modeling Time series prediction |
| gptkbp:bfsParent |
gptkb:LSTM
|
| gptkbp:bfsLayer |
7
|