Learning Model Creation With Blockchain Data Usage Tracking

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing machine learning technologies fail to track the usage aspect of data, particularly medical data, leading to unintended usage and disclosure of unnecessary information.

Innovation Solution

A learning model creation device and method that includes a learning request reception unit, data readout unit, and learning model creation unit, which designates data usage and creates models while preserving learning processes on a blockchain, enabling data usage tracking and preventing unintended disclosure.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If data is provided to multiple learning models multiple times, then the accuracy of learning models is improved, but the ability to grasp data usage aspects is worsened

Engineering Contradiction:
Improvelearning model accuracyVSAvoiddata usage aspect information
Core Design Contradiction:
Measurement precisionVSLoss of information

Solution Approach 1:

The patent implements a feedback mechanism where the management server receives notifications from learning models about data usage, tracks usage counts, and manages data provision based on predetermined thresholds. This feedback loop enables the system to maintain accurate tracking of data usage aspects even when data is reused multiple times across different learning models, thereby resolving the contradiction between improving model accuracy through data reuse and maintaining visibility into data usage patterns.

Inventive Principle:
Principle #23Feedback

2Productivity

If data is reused multiple times for creating multiple learning models, then the productivity of machine learning is improved, but the control over data usage is worsened

Engineering Contradiction:
Improvemachine learning efficiencyVSAvoiddata usage control
Core Design Contradiction:
ProductivityVSEase of operation

Solution Approach 1:

The patent introduces a management server as an intermediary between data providers and learning models. This intermediary automatically tracks data usage, manages provision rights, and enforces predetermined conditions for data reuse. By delegating these control functions to the management server, the system enables high-productivity data reuse while maintaining automated control over usage policies, thus resolving the contradiction between productivity improvement and usage control.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Adaptability or versatility

If various types of information are included in data, then the versatility of learning models is improved, but the risk of unintended disclosure is worsened

Engineering Contradiction:
Improvelearning model versatilityVSAvoidunintended disclosure risk
Core Design Contradiction:
Adaptability or versatilityVSObject-affected harmful factors

Solution Approach 1:

The patent implements dynamic control over data provision where the management server adjusts data access based on predetermined conditions and usage history. The system can dynamically restrict or allow access to specific data elements based on the learning model's purpose, usage patterns, and predetermined policies. This dynamic approach enables versatile learning models to access comprehensive data when needed while automatically preventing unintended disclosure through condition-based access control.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP4693316A1Learning model creation device and learning model creation method
Publication Date: 2026.02.11 FUJIFILM CORP
  • EP4693316A1 patent drawingFigure 1~2
  • EP4693316A1 patent drawingFigure 3~4
  • EP4693316A1 patent drawingFigure 5

AI summary

Provided are a learning model creation device and a learning model creation method that can grasp a usage aspect of data. A processor (40) receives a learning request and starts a process of creating a learning model. In a case where the process is started, the processor (40) reads out medical data corresponding to a range designated in the learning request from a medical data management server (22). Then, the processor (40) creates the learning model using the read out medical data and a learning algorithm designated in the learning request.