Determination device, determination method, and determination program

The determining device addresses the challenge of generating teacher data by assessing uncertainty in machine learning models, allowing for pre-judgment on accuracy improvements and reducing unnecessary experiments.

JP2025037405APending Publication Date: 2025-03-18HITACHI LTD
1 Cites 0 Cited by

Patent Information

Application Number
JP2023144321
Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Filing Date
2023-09-06
Publication Date
2025-03-18

AI Technical Summary

Technical Problem

Existing machine learning technologies do not adequately address the need to generate teacher data before using it for uncertainty calculation, leading to inaccurate high-precision uncertainty assessments based solely on generation conditions before teacher data is generated.

Method used

A determining device that accesses a database storing first generation data related to generation results, determines uncertainty regarding determination success or failure, and decides whether generating a second generation result differs from the first based on this uncertainty.

Benefits of technology

Enables a pre-judgment on whether the accuracy of a machine learning model improves, allowing for the determination of necessary experiments and reducing unnecessary costs and time by assessing uncertainty effectively.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

To provide, in advance, a result of determination as to whether or not the accuracy of a machine learning model can be improved.SOLUTION: A determination device can access a database storing one or more sets of combinations of first generated data relating to a first generated result generated under a first generation condition and determination success or failure of the first generated result, and executes: acquisition processing for acquiring the first generated data from the database; determination processing for outputting determination uncertainty relating to the determination success or failure of the first generated result by inputting the first generated data acquired by the acquisition processing into a determination model that is trained on the basis of the sets and outputs the determination uncertainty relating to the determination success or failure of the first generated result upon receiving the first generated data; and decision processing for deciding as to whether there is a need to generate a second generated result different from the first generated result on the basis of the determination uncertainty output by the determination processing.SELECTED DRAWING: Figure 2
Need to check novelty before this filing date? Find Prior Art