FIELD: computing technology.
SUBSTANCE: invention relates to computer systems based on biological models and can be used in neurocomputers, technical systems based on neural networks, in pattern recognition, image analysis and processing, and in artificial intelligence. The technical result of the claimed solution is achieved by the neuron model being based on dendritic computing and including information inputs; information output; dendrites, the synapses whereof form clusters; and a compute core for forming the output neuron signal, an input for a feedback signal is also provided, by means whereof a neuron effectiveness signal is supplied to the dendrites in order to account for the effectiveness thereof.
EFFECT: selective recognition of input objects with a possibility of automatically creating most informative object-recognising syndrome patterns.
3 cl, 4 dwg
Title | Year | Author | Number |
---|---|---|---|
NEAR-REAL IMPULSE NEURON | 2015 |
|
RU2598298C2 |
SINGLE-LAYER PERCEPTRON BASED ON SELECTIVE NEURONS | 2015 |
|
RU2597497C2 |
SINGLE-LAYER PERCEPTRON, SIMULATING REAL PERCEPTRON PROPERTIES | 2015 |
|
RU2597496C1 |
NEURON SIMULATION OF REAL NEURON PROPERTIES | 2014 |
|
RU2597495C2 |
DEVICE FOR SIMULATING NEURON | 1991 |
|
RU2028669C1 |
NEURON SIMULATOR | 0 |
|
SU1501101A1 |
DEVICE FOR MODELING DEVELOPMENT OF NEURON MORPHOLOGY | 0 |
|
SU1815658A1 |
METHOD AND DEVICE FOR NEURAL TIME CODING, TRAINING AND RECOGNITION | 2012 |
|
RU2597504C2 |
DEVICE FOR MODELING NEURON OF HIGHER ORDER | 0 |
|
SU1561076A1 |
DEVICE FOR MODELING NEURON | 0 |
|
SU1585811A1 |
Authors
Dates
2022-08-01—Published
2021-04-29—Filed