Learning Algorithm Access Control for Authorized Model Training

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing systems face challenges in managing access and usage of learning algorithms and models, particularly in industrial facilities, leading to inefficiencies and unauthorized use, which hampers the development and deployment of tailored machine learning solutions.

Innovation Solution

A system that includes a storage unit for learning algorithms with user authority management, a control unit for restricting access, and a learning processing unit that performs learning using authorized data, allowing for the registration of new algorithms and models while ensuring proper usage and efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If access to learning algorithms is restricted to authorized users only, then security and proper usage are improved, but system complexity and access management overhead increase

Engineering Contradiction:

Improveauthorized use of learning algorithmsVSAvoidaccess management system

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The system segments access rights by creating distinct account types (developer accounts and user accounts) with different authorization levels. Each learning algorithm has associated use authorities that are selectively assigned to specific accounts, enabling fine-grained control over who can use which algorithms without requiring a monolithic access control system.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system performs preliminary authorization checks before allowing access to learning algorithms. The use control unit verifies whether a user account has the necessary use authority for a selected learning algorithm before permitting its execution, preventing unauthorized use in advance rather than reacting to violations afterward.

Inventive Principle:

Principle #10Preliminary action

2Adaptability or versatility

If multiple learning algorithms are made available to users, then system versatility and functionality are improved, but difficulty in managing and controlling access increases

Engineering Contradiction:

Improveavailability of learning algorithmsVSAvoidalgorithm management system

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The learning processing unit is designed to handle multiple different learning algorithms through a unified interface. The system can execute various learning algorithms (first learning algorithm, second learning algorithm, etc.) using the same core processing infrastructure, allowing versatility without proportionally increasing management complexity.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Solution Approach 2:

The use control unit acts as an intermediary between user accounts and learning algorithms. It manages the complex relationships between multiple users and multiple algorithms by maintaining use authority records and routing requests appropriately, shielding users from the underlying complexity of algorithm management.

Inventive Principle:

Principle #24Intermediary (Mediator)

3Reliability

If learning algorithms are restricted to authorized accounts, then unauthorized use is prevented, but ease of operation and user accessibility decrease

Engineering Contradiction:

Improvecontrol over learning algorithm usageVSAvoiduser access to learning algorithms

Core Design Contradiction:

ReliabilityVSEase of operation

Solution Approach 1:

Developer accounts can automatically grant use authority to user accounts for learning algorithms they have developed. This self-service mechanism allows authorized users to access algorithms without manual approval, maintaining ease of operation while ensuring proper authorization. The system automatically verifies and permits access when authorization exists.

Inventive Principle:

Principle #25Self-service

4Adaptability or versatility

If new learning algorithms are registered and made available, then system functionality and versatility are improved, but management complexity and access control difficulty increase

Engineering Contradiction:

Improvenew learning algorithmsVSAvoidregistration and management system

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

When a new learning algorithm is registered, the system automatically creates the associated use authority structure and links it to the developer's account. This preliminary setup ensures that access control mechanisms are in place before the algorithm becomes available to users, simplifying subsequent management rather than requiring retroactive configuration.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS12563044B2Apparatus, method, and computer readable medium

Publication Date: 2026.02.24 YOKOGAWA ELECTRIC CORP

AI summary

Provided is an apparatus including: a storage unit configured to store an account having a use authority for each of a plurality of learning algorithms; a use control unit configured to restrict use of each learning algorithm to a user of an account having a use authority for the learning algorithm; and a learning processing unit configured to perform learning processing of a learning model by one learning algorithm of the plurality of learning algorithms by using learning data supplied from a user of an account having a use authority for the one learning algorithm.