Title:
NEURAL NETWORK DEVICE
Document Type and Number:
Japanese Patent JP3178884
Kind Code:
B2
Abstract:
PURPOSE: To provide the neural network device which does not require a reverse propagation signal for learning and can structure a large-scale neural network with inexpensive circuits.
CONSTITUTION: This neural network device is equipped with a synapses 200 having a random number generator 210 which statistically has partiality or no partially, has its operation start and stop and the holding and initialization of its internal state controlled according to an external control signal, and holds and outputs a random number value according to the external control signal and a load value varying means 220 which varies its load value according to the held random number value, a comparing means 300 which compares a neuron output as an external output signal among neuron outputs with an external tutor signal, and a verifying means 400 which verifies the signal outputted from the comparing means 300 and outputs a signal for controlling whether the load value is altered or restored.
More Like This:
WO/2012/027673 | POLICY CONFLICT CLASSIFIER |
JP2000066703 | SELF BIAS CONDITIONING DEVICE |
JPH04292713 | HEATING COOKER |
Inventors:
Ken Shima
Application Number:
JP7448892A
Publication Date:
June 25, 2001
Filing Date:
March 30, 1992
Export Citation:
Assignee:
Toshiba Corporation
International Classes:
G06F15/18; G06G7/60; G06N3/04; G06N3/063; G06N3/08; G06N99/00; (IPC1-7): G06G7/60; G06F15/18
Domestic Patent References:
JP1243169A | ||||
JP2238561A | ||||
JP3242780A |
Attorney, Agent or Firm:
Takehiko Suzue