# US 2020/0394550 A1 · Quantum Circuit Learning Device, Quantum Circuit Learning Method, and Recording Medium

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

- Number: US 2020/0394550 A1
- Document ID: US20200394550A1
- Status: Published application
- Kind: A1
- Application number: 17/002,419
- Filed: 2020-08-25
- Published: 2020-12-17
- Earliest claimed: 2018-02-26
- Assignees: Kyoto University; Osaka University
- Applicants: Kyoto University; Osaka University
- Inventors: Keisuke Fujii; Makoto Negoro; Kosuke Mitarai; Masahiro Kitagawa
- Web page: https://patentgrep.com/patent/US20200394550A1

## CPC

- G06N 10/60: Quantum algorithms, e.g. based on quantum optimisation, quantum Fourier or Hadamard transforms
- G06F 18/214: Generating training patterns; Bootstrap methods, e.g. bagging or boosting
- G06F 18/2185: {the supervisor being an automated module, e.g. intelligent oracle}
- G06N 10/20: Models of quantum computing, e.g. quantum circuits or universal quantum computers
- G06N 20/00: Machine learning
- H03K 19/195: using superconductive devices
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- H03K: PULSE TECHNIQUE (measuring pulse characteristics G01R; modulating sinusoidal oscillations with pulses H03C; transmission of digital information H04L; discriminator circuits detecting phase difference between two signals by counting or integrating cycles of oscillation H03D3/04; automatic control, starting, synchronisation or stabilisation of generators of electronic oscillations or pulses where the type of generator is irrelevant or unspecified H03L; coding, decoding or code conversion, in general H03M)

## Abstract

The quantum circuit learning device includes a signal input unit that provides a quantum circuit including plural quantum bits with an input signal, a signal acquisition unit that observes states of quantum bits that the quantum circuit includes and acquires an output signal based on the observed states, and an adjustment unit that adjusts a circuit parameter that defines a circuit configuration of the quantum circuit, using an output signal that the signal acquisition unit acquires and a cost function that is set based on a teacher signal corresponding to the output signal.

## Claims

### Claim 1

A quantum circuit learning device, comprising:
a memory; and
a processor coupled to the memory, the processor being configured to perform a process comprising:
providing a quantum circuit U({θk}), which includes a plurality of quantum bits, with an input signal x(j), wherein θk is a circuit parameter that defines a circuit configuration of the quantum circuit U({θk}), k is an index of the circuit parameter, and j is an index of the signal;
observing states of the quantum bits included in the quantum circuit U({θk}) and acquiring a value <A(x(j), {θk})> that is calculated from an output of the quantum circuit U({θk}) based on the observed states;
adjusting the circuit parameter θk, using the acquired value <A(x(j), {θk})> and using a cost function that is set based on a teacher signal y(j) corresponding to the value <A(x(j), {θk})>; and
at a time of adjusting an l-th circuit parameter θ1 in the circuit parameter θ, calculating a gradient in accordance with the following Formula (1), the gradient being calculated from a difference between a value <A(x(j), {θ1, . . . , θ1+ϵ, θ1+1, . . . })> calculated from an output of the quantum circuit U({θk}) in a case in which the l-th circuit parameter θ1 is increased by a set value ϵ, and a value <A(x(j), {θ1, . . . , θ1−ϵ, θ1+1, . . . })> calculated from an output of the quantum circuit U({θk}) in a case in which the l-th circuit parameter θ1 is decreased by the set value ϵ, and adjusting the circuit parameter θ1 in such a way as to minimize the cost function by means of a gradient method, using the calculated gradient value:
∂
∂
θ
l

〈
A

(
x
(
j
)
,
{
θ
k
}
)
〉
=
1
2

sin

(
ϵ
)

(
〈
A

(
x
(
j
)
,
{
θ
1
,
…


θ
l
+
ϵ
,
θ
l
+
1
,
…

}
)
〉
-
〈
A

(
x
(
j
)
,
{
θ
1
,
…


θ
l
-
ϵ
,
θ
l
+
1
,
…

}
)
〉
)
[
Formula


1
]

### Claim 2 (depends on 1)

The quantum circuit learning device according to claim 1, wherein the circuit parameter includes a parameter for controlling an operator operating on the quantum bits.

### Claim 3

A quantum circuit learning method that causes a computer to execute a process comprising:
by a processor,
providing a quantum circuit U({θk}), which includes a plurality of quantum bits, with an input signal x(j), wherein θk is a circuit parameter that defines a circuit configuration of the quantum circuit U({θk}), k is an index of the circuit parameter, and j is an index of the signal;
observing states of the quantum bits included in the quantum circuit U({θk}) and acquiring a value <A(x(j), {θk})> that is calculated from an output of the quantum circuit U({θk}) based on the observed states;
adjusting the circuit parameter θk, using the acquired value <A(x(j), {θk})> and using a cost function that is set based on a teacher signal y(j) corresponding to the value <A(x(j), {θk})>; and
at a time of adjusting an l-th circuit parameter θ1 in the circuit parameter θ, calculating a gradient in accordance with the following Formula (1), the gradient being calculated from a difference between a value <A(x(j), {θ1, . . . , θ1+ϵ, θ1+1, . . . })> calculated from an output of the quantum circuit U({θk}) in a case in which the l-th circuit parameter θ1 is increased by a set value ϵ, and a value <A(x(j), {θ1, . . . , θ1−ϵ, θ1+1, . . . })> calculated from an output of the quantum circuit U({θk}) in a case in which the l-th circuit parameter 0i is decreased by the set value ϵ, and adjusting the circuit parameter θ1 in such a way as to minimize the cost function by means of a gradient method, using the calculated gradient value:
∂
∂
θ
l

〈
A

(
x
(
j
)
,
{
θ
k
}
)
〉
=
1
2

sin

(
ϵ
)

(
〈
A

(
x
(
j
)
,
{
θ
1
,
…


θ
l
+
ϵ
,
θ
l
+
1
,
…

}
)
〉
-
〈
A

(
x
(
j
)
,
{
θ
1
,
…


θ
l
-
ϵ
,
θ
l
+
1
,
…

}
)
〉
)
.
[
Formula


1
]

### Claim 4

A non-transitory recording medium storing a computer program that causes a computer to execute a process comprising:
providing a quantum circuit U({θk}), which includes a plurality of quantum bits, with an input signal x(j), wherein θk is a circuit parameter that defines a circuit configuration of the quantum circuit U({θk}), k is an index of the circuit parameter, and j is an index of the signal;
observing states of the quantum bits included in the quantum circuit U({θk}) and acquiring a value <A(x(j), {θk})> that is calculated from an output of the quantum circuit U({θk}) based on the observed states;
adjusting the circuit parameter θk, using the acquired value <A(x(j), {θk})> and using a cost function that is set based on a teacher signal y(j) corresponding to the value <A(x(j), {θk})>, and
at a time of adjusting an l-th circuit parameter θ1 in the circuit parameter θ, calculating a gradient in accordance with the following Formula (1), the gradient being calculated from a difference between a value <A(x(j), {θ1, . . . , θ1+ϵ, θ1+1, . . . })> calculated from an output of the quantum circuit U({0k}) in a case in which the l-th circuit parameter θ1 is increased by a set value ϵ, and a value <A(x(j), {θ1, . . . , θ1−ϵ, θ1+1, . . . })> calculated from an output of the quantum circuit U({θk}) in a case in which the l-th circuit parameter θ1 is decreased by the set value ϵ, and adjusting the circuit parameter θ1 in such a way as to minimize the cost function by means of a gradient method, using the calculated gradient value:
∂
∂
θ
l

〈
A

(
x
(
j
)
,
{
θ
k
}
)
〉
=
1
2

sin

(
ϵ
)

(
〈
A

(
x
(
j
)
,
{
θ
1
,
…


θ
l
+
ϵ
,
θ
l
+
1
,
…

}
)
〉
-
〈
A

(
x
(
j
)
,
{
θ
1
,
…


θ
l
-
ϵ
,
θ
l
+
1
,
…

}
)
〉
)
[
Formula


1
]
