Similarity determination method, similarity calculation unit, diffusion-type learning network, and executive program of neural network
AssigneeNTT, Inc.
InventorTakao YAMASHITA
Life of this patent
Estimating the term…
Abstract
A value obtained by dividing the number of inputs in which both wi and yi are the value H by a value obtained by adding the number of inputs in which yi is the value H to the number of inputs in which wi is the value His calculated and output as similarity representing the degree of similarity.
Claims14
A similarity determination method for calculating a degree of similarity between an input of a learning phase and an input of a similarity determination phase using a perceptron obtained by modeling a nerve cell, the similarity determination method comprising:
- receiving one or more input values, wherein
- one of a value L and a value His input to each input value,
- when an i-th input value in the learning phase is represented as xi and
- an i-th input value in the similarity determination phase is represented as yi,
- wi is assigned to the i-th input value,
- one of the value L and the value His set to the value wi;
- in the learning phase, setting the value wi of a weight assigned to the i-th input value to the value of xi;
- and
- in the similarity determination phase,
- calculating a value obtained by dividing a number of inputs in which both wi and yi are the value H by a value obtained by adding a number of inputs in which yi is the value H to a number of inputs in which the value of wi is the value H as similarity representing the degree of similarity.
A similarity determination method for calculating a degree of similarity between an input of a learning phase and an input of a similarity determination phase using a perceptron obtained by modeling a nerve cell, the similarity determination method comprising:
- in the similarity determination phase, calculating a divisive normalization similarity using a formula that incorporates an operation caused by a phenomenon called a shunt effect of a nerve cell into a model of the perceptron.
- [
- Formula
-
- 6
- ]
-
- 2
-
- ∑
- j
- =
- 1
- 6
- y
- j
- ·
- w
- j
-
- x
-
- 2
- +
- ∑
- j
- =
- 1
- 6
- y
- j
- =
- 2
-
- (
- y
- ·
- w
- )
-
- w
- i
-
- 2
- +
-
- y
- i
-
- 2
A diffusive learning network in which a plurality of similarity calculators having some or all of inputs with respect to a plurality of inputs are connected, and further, outputs of the respective similarity calculators are input to a perceptron, wherein
- the similarity calculator
- receives one or more input values,
- one of a value L and a value His input to each input value,
- when a value of an i-th input in a learning phase is represented as xi, and
- a value of an i-th input in a similarity determination phase is represented as yi,
- a value wi is assigned to the i-th input,
- one of the value L and the value H is set to the value wi,
- in the learning phase, the similarity calculator sets the value wi of a weight assigned to the i-th input to the value of xi, and
- in the similarity determination phase,
- the similarity calculator performs similarity calculation that calculates a value obtained by dividing a number of inputs in which both wi and yi are the value H by a value obtained by adding a number of inputs in which the value of yi is the value H to a number of inputs in which the value of wi is the value H as similarity representing a degree of similarity.
(canceled)
Dependencies are inferred from the claim text.