# US 2021/0279559 A1 · Spiking neural network device and learning method of spiking neural network device

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2021/0279559 A1
- Document ID: US20210279559A1
- Status: Published application
- Kind: A1
- Application number: 17/007,331
- Filed: 2020-08-31
- Published: 2021-09-09
- Earliest claimed: 2020-03-03
- Assignees: Kabushiki Kaisha Toshiba
- Applicants: Kabushiki Kaisha Toshiba
- Inventors: Yoshifumi NISHI; Kumiko NOMURA; Takao MARUKAME; Koichi MIZUSHIMA
- Web page: https://patentgrep.com/patent/US20210279559A1

## CPC

- G06N 3/049: Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
- G06N 3/063: using electronic means
- G06N 3/047: Probabilistic or stochastic networks
- G06N 3/0495: Quantised networks; Sparse networks; Compressed networks
- G06N 3/0499: Feedforward networks
- G06N 3/065: Analogue means
- G06N 3/08: Learning methods
- G06N 3/088: Non-supervised learning, e.g. competitive learning
- G06N 3/09: Supervised learning
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

A spiking neural network device according to an embodiment includes a synaptic element, a neuron circuit, a determinator, a synaptic depressor, and a synaptic potentiator. The synaptic element has a variable weight and outputs, in response to input of a first spike signal, a synaptic signal having intensity adjusted in accordance with the weight. The neuron circuit outputs a second spike signal in a case where the synaptic signal is inputted and a predetermined firing condition for the synaptic signal is satisfied. The determinator determines whether or not the weight is to be updated on a basis of an output frequency of the second spike signal by the neuron circuit. The synaptic depressor performs depression operation for depressing the weight in a case where it is determined that the weight is to be updated. The synaptic potentiator performs potentiating operation for potentiating the weight.

## Claims

### Claim 1

A spiking neural network device comprising:
a synaptic element configured to have a variable weight, and configured to output, in response to input of a first spike signal, a synaptic signal whose intensity is adjusted in accordance with the weight;
a neuron circuit configured to output a second spike signal in a case where the synaptic signal is inputted and a predetermined firing condition for the synaptic signal is satisfied;
a determinator configured to determine whether or not the weight is to be updated on a basis of an output frequency of the second spike signal by the neuron circuit;
a synaptic depressor configured to perform depression operation for depressing the weight in a case where it is determined that the weight is to be updated; and
a synaptic potentiator configured to perform potentiating operation for potentiating the weight.

### Claim 2 (depends on 1)

The spiking neural network device according to claim 1,
wherein the determinator determines that the weight is to be updated in a case where the output frequency is equal to or greater than a threshold.

### Claim 3 (depends on 1)

The spiking neural network device according to claim 1,
wherein the output frequency is represented by a variable that increases each time the neuron circuit outputs the second spike signal and decreases with time.

### Claim 4 (depends on 1)

The spiking neural network device according to claim 1,
wherein the synaptic depressor performs the depression operation in a case where it is determined that the weight is to be updated and the neuron circuit outputs the second spike signal.

### Claim 5 (depends on 1)

The spiking neural network device according to claim 1,
wherein the synaptic potentiator performs the potentiating operation in a case where a predetermined potentiating condition is satisfied.

### Claim 6 (depends on 5)

The spiking neural network device according to claim 5,
wherein the synaptic potentiator performs the potentiating operation in a case where the determinator determines that the weight is to be updated and the potentiating condition is satisfied.

### Claim 7 (depends on 1)

The spiking neural network device according to claim 1,
wherein the synaptic element includes a storage element for storing a value of the weight by a resistance value and a metal oxide semiconductor (MOS) transistor.

### Claim 8 (depends on 7)

The spiking neural network device according to claim 7,
wherein the storage element has a plurality of resistance states and transition between the plurality of resistance states is stochastic.

### Claim 9 (depends on 7)

The spiking neural network device according to claim 7,
wherein the synaptic element further includes a diode for connecting the MOS transistor and the storage element to each other.

### Claim 10

A learning method of a spiking neural network device,
the spiking neural network device including
a synaptic element configured to have a variable weight, and configured to output, in response to input of a first spike signal, a synaptic signal whose intensity is adjusted in accordance with the weight, and
a neuron circuit configured to output a second spike signal in a case where the synaptic signal is inputted and a predetermined firing condition for the synaptic signal is satisfied,
the learning method comprising:
a determination step of determining whether or not the weight is to be updated on a basis of an output frequency of the second spike signal by the neuron circuit;
a synapse depressing step of performing depression operation for depressing the weight in a case where it is determined that the weight is to be updated; and
a synapse potentiating step of performing potentiating operation for potentiating the weight.
