# US 2025/0094260 A1 · Information processing device, information processing method, and computer program product

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

- Number: US 2025/0094260 A1
- Document ID: US20250094260A1
- Status: Published application
- Kind: A1
- Application number: 18/664,411
- Filed: 2024-05-15
- Published: 2025-03-20
- Earliest claimed: 2023-09-15
- Assignees: Kabushiki Kaisha Toshiba
- Applicants: Kabushiki Kaisha Toshiba
- Inventors: Akihiro YAMAGUCHI; Ken UENO
- Web page: https://patentgrep.com/patent/US20250094260A1

## CPC

- G06F 11/0751: {Error or fault detection not based on redundancy (power supply failures G06F1/30; network fault management H04L41/06)}
- G06N 20/00: Machine learning
- G06N 3/044: Recurrent networks, e.g. Hopfield networks
- G06N 3/045: Combinations of networks
- G06N 3/08: Learning methods
- G06N 3/084: Backpropagation, e.g. using gradient descent
- G06N 5/01: Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
- G06N 7/01: Probabilistic graphical models, e.g. probabilistic networks
- G06N 20/10: using kernel methods, e.g. support vector machines [SVM]
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

An information processing device includes one or more processors. The one or more processors are configured to: adjust a first distance between first time-series data and a plurality of identification patterns corresponding to a plurality of classes by using a degree of existence indicating a degree by which a plurality of times of the first time-series data are times at which the plurality of respective identification patterns exist; calculate a first feature amount indicating a feature of the first time-series data by using the adjusted first distance; and obtain, by inputting the first feature amount to an estimation model learning a parameter, the similar identification pattern, and the degree of existence to reduce an estimation error, the identification pattern similar to the first time-series data, the class, and the degree of existence.

## Claims

### Claim 1

An information processing device comprising:
one or more processors configured to:
adjust a first distance between input first time-series data and a plurality of identification patterns respectively corresponding to a plurality of classes by using a degree of existence indicating a degree by which a plurality of times of the first time-series data are times at which the plurality of respective identification patterns exist;
calculate a first feature amount indicating a feature of the first time-series data by using the adjusted first distance; and
obtain an identification pattern similar to the first time-series data, the class, and the degree of existence by inputting the first feature amount to an estimation model, the estimation model inputting a second feature amount indicating a feature of second time-series data for learning to estimate a class of the classes corresponding to an identification pattern similar to the second time-series data among the plurality of identification patterns, the estimation model learning a parameter, the similar identification pattern, and the degree of existence indicating a time at which the similar identification pattern exists to reduce an estimation error.

### Claim 2 (depends on 1)

The device according to claim 1, wherein
the one or more processors are configured to execute a learning process of updating the parameter of the estimation model, the plurality of identification patterns, and the degree of existence by using the second time-series data to reduce an error between the class estimated by the estimation model and a correct answer class.

### Claim 3 (depends on 2)

The device according to claim 2, wherein
the one or more processors are configured to update the degrees of existence such that values of a plurality of the degrees of existence at adjacent times come close to each other.

### Claim 4 (depends on 1)

The device according to claim 1, wherein
the one or more processors are configured to adjust the first distance by multiplying by the degree of existence.

### Claim 5 (depends on 1)

The device according to claim 1, wherein
the first distance is a sum of distances between the first time-series data and one or more of the plurality of identification patterns.

### Claim 6 (depends on 1)

The device according to claim 1, wherein
the one or more processors are configured to display, on a display device, the class estimated for the first feature amount, the identification pattern corresponding to the estimated class, and the degree of existence.

### Claim 7 (depends on 1)

The device according to claim 1, wherein
the first distance is calculated based on a plurality of types of distances between the first time-series data and the plurality of identification patterns and a contribution degree of each of the plurality of types of distances, and
the estimation model learns the parameter together with the similar identification pattern, the degree of existence indicating the time at which the similar identification pattern exists, and the contribution degree to reduce an estimation error.

### Claim 8 (depends on 7)

The device according to claim 7, wherein
the plurality of types of distances include two or more of:
a difference between a statistic of partial time-series data corresponding to a length of the identification pattern included in the first time-series data and a statistic of the identification pattern;
a distance between a shape of the partial time-series data and a shape of the identification pattern;
a distance between a shape of the normalized partial time-series data and a shape of the identification pattern;
a distance between a frequency component of the partial time-series data and a frequency component of the identification pattern; and
a distance between a p-th (p is an integer of 1 or more) derivative of the partial time-series data and a p-th derivative of the identification pattern.

### Claim 9 (depends on 7)

The device according to claim 7, wherein
the one or more processors are configured to display, on a display device, the class estimated for the first feature amount, the identification pattern corresponding to the estimated class, the degree of existence, and the contribution degree.

### Claim 10 (depends on 1)

The device according to claim 1, wherein
the one or more processors are configured to:
adjust the first distance by using the degree of existence;
calculate the first feature amount; and
obtain the identification pattern similar to the first time-series data, the class, and the degree of existence by inputting the first feature amount to the estimation model.

### Claim 11

An information processing method executed by the information processing device, the method comprising:
adjusting a first distance between input first time-series data and a plurality of identification patterns respectively corresponding to a plurality of classes by using a degree of existence indicating a degree by which a plurality of times of the first time-series data are times at which the plurality of respective identification patterns exist;
calculating a first feature amount indicating a feature of the first time-series data by using the adjusted first distance; and
obtaining an identification pattern similar to the first time-series data, the class, and the degree of existence by inputting the first feature amount to an estimation model, the estimation model inputting a second feature amount indicating a feature of second time-series data for learning to estimate a class of the classes corresponding to an identification pattern similar to the second time-series data among the plurality of identification patterns, the estimation model learning a parameter, the similar identification pattern, and the degree of existence indicating a time at which the similar identification pattern exists to reduce an estimation error.

### Claim 12

A computer program product comprising a non-transitory computer-readable medium including programmed instructions, the instructions causing a computer to execute:
adjusting a first distance between input first time-series data and a plurality of identification patterns respectively corresponding to a plurality of classes by using a degree of existence indicating a degree by which a plurality of times of the first time-series data are times at which the plurality of respective identification patterns exist;
calculating a first feature amount indicating a feature of the first time-series data by using the adjusted first distance; and
obtaining an identification pattern similar to the first time-series data, the class, and the degree of existence by inputting the first feature amount to an estimation model, the estimation model inputting a second feature amount indicating a feature of second time-series data for learning to estimate a class of the classes corresponding to an identification pattern similar to the second time-series data among the plurality of identification patterns, the estimation model learning a parameter, the similar identification pattern, and the degree of existence indicating a time at which the similar identification pattern exists to reduce an estimation error.
