Back to results
US 2019/0095783 A1Published application
Deep learning apparatus for ann having pipeline architecture
AssigneeKomosystem Co., Ltd.
InventorLokwon KIM
Life of this patent
Estimating the term…
Abstract
A deep learning apparatus for an artificial neural network (ANN) having pipeline architecture. The deep learning apparatus for an ANN simultaneously performs output value processing, corrected input data processing, corrected output value processing, weight correction, input bias correction, and output bias correction using pipeline architecture, thereby reducing calculation time for learning and reducing required memory capacity.
Claims11
A deep learning apparatus for an artificial neural network (ANN), comprising:
- an input memory storing an input batch therein, the input batch being an assembly of input cases, each of which is comprised of a plurality of pieces of input data;
- a data processing module comprising a plurality of stages connected to each other, each of the stages being comprised of an combinational logic circuit and a register to correct weights between input nodes and hidden nodes using the input batch;
- and
- a controller generating a control signal, by which the pieces of input data of the input cases are input to the data processing module, or a control signal, by which an operation of each of the stages is controlled.
1 more element
Dependencies are inferred from the claim text.