Computer system, and method and program for controlling edge device

• US 10,460,258 B2
• Filed: 11/30/2016
• Issued: 10/29/2019
• Est. Priority Date: 11/30/2016
• Status: Active Grant

• Current Assignee: Optim LLC
• Original Assignee: Optim LLC
• Inventors: Sugaya, Shunji
• Primary Examiner(s): Chen, Alan
• Application Number:
  US15/568,852
• Publication Number:
  US 20180293513A1
• Time in Patent Office: 1,063 Days
• Field of Search: None
• CPC Class Codes: G06F 9/50 Allocation of resources, e.g. of the central processing unit [CPU]
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: H04L 12/66 Arrangements for connecting between networks having differing types of switching systems, e.g. gateways
• CPC Class Codes: H04L 67/125 involving control of end-device applications over a network

Abstract

The present invention is to provide a computer system and a method and a program for controlling an edge device that acquire data for a predetermined machine learning from a combination of sensors in a network, perform possible learning with the sensors without the user'"'"'s intention, and output the result. According to the present invention, a computer system that performs machine learning with an edge device 100 connected with a gateway 200 detects an edge device 100 connected with the gateway 200, determines the combination of the detected edge devices 100, determines a program for an edge device 100 and a program for machine learning based on the determined combination, and causes the edge devices 100 to execute the program for an edge device 100 and causes a predetermined computer to execute the program for machine learning.

Claims

1. A computer system, comprising:

a communication device; and

a processor that;

acquires device data on edge devices connected with a gateway via the communication device;

determines a type corresponding to a combination of the edge devices from among a plurality of types, based on the acquired device data;

determines a first program for the edge devices based on the determined type;

transmits, via the communication device, data on the first program to the edge devices to execute the first program;

receives, via the communication device, device result data from each of the edge devices, the device result data being data which is acquired by each of the edge devices based on the first program; and

transmits, via the communication device, teacher data corresponding to the device result data to a first edge device whose performance meets or exceeds a predetermined performance among the edge devices, thereby causing the first edge device to perform machine learning for the device result data by using the teacher data.

2. The computer system according to claim 1, wherein the processor:

determines a second program for machine learning based on the determined type; and

transmits data on the second program to the first edge device when transmitting the teacher data to the first edge device.

3. The computer system according to claim 1, wherein the processor:

determines a second program for machine learning based on the determined type;

transmits, via the communication device, data on the second program to a predetermined computer to execute the second program; and

when a second edge device among the edge devices is less than the predetermined performance, transmits the device result data of the second edge device and the teacher data corresponding to the device result data to the predetermined computer, thereby causing the predetermined computer to perform machine learning for the device result data by using the teacher data.

4. A method executed by a computer system, comprising:

acquiring device data on edge devices connected with a gateway;

determining a type corresponding to a combination of the edge devices from among a plurality of types, based on the acquired device data;

determining a first program for the edge devices based on the determined type;

transmitting data on the first program to the edge devices to execute the first program;

receiving device result data from each of the edge devices, the device result data being data which is acquired by each of the edge devices based on the first program; and

transmitting teacher data corresponding to the device result data to a first edge device whose performance meets or exceeds a predetermined performance among the edge devices, thereby causing the first edge device to perform machine learning for the device result data by using the teacher data.

5. The method according to claim 4, further comprising:

determining a second program for machine learning based on the determined type; and

transmitting data on the second program to the first edge device when transmitting the teacher data to the first edge device.

6. The method according to claim 4, further comprising:

determining a second program for machine learning based on the determined type;

transmitting data on the second program to a predetermined computer to execute the second program; and

when a second edge device among the edge devices is less than the predetermined performance, transmitting the device result data of the second edge device and the teacher data corresponding to the device result data to the predetermined computer, thereby causing the predetermined computer to perform machine learning for the device result data by using the teacher data.

7. A non-transitory computer-readable recording medium having a set of instructions for causing a computer system to execute:

acquiring device data on edge devices connected with a gateway;

determining a type corresponding to a combination of the edge devices from among a plurality of types, based on the acquired device data;

determining a first program for the edge devices based on the determined type;

transmitting data on the first program to the edge devices to execute the first program;

receiving device result data from each of the edge devices, the device result data being data which is acquired by each of the edge devices based on the first program; and

transmitting teacher data corresponding to the device result data to a first edge device whose performance meets or exceeds a predetermined performance among the edge devices, thereby causing the first edge device to perform machine learning for the device result data by using the teacher data.

8. The non-transitory computer-readable recording medium according to claim 7, wherein the set of instructions causes the computer system to further execute:

determining a second program for machine learning based on the determined type; and

transmitting data on the second program to the first edge device when transmitting the teacher data to the first edge device.

9. The non-transitory computer-readable recording medium according to claim 7, wherein the set of instructions causes the computer system to further execute:

determining a second program for machine learning based on the determined type;

transmitting data on the second program to a predetermined computer to execute the second program; and

when a second edge device among the edge devices is less than the predetermined performance, transmitting the device result data of the second edge device and the teacher data corresponding to the device result data to the predetermined computer, thereby causing the predetermined computer to perform machine learning for the device result data by using the teacher data.