Learning Model and Dataset Management with Update Notifications
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The proliferation of learning models and data created by general users makes it difficult for users to grasp the status and suitability of these models and data for their purposes, and existing methods increase operational costs and fail to effectively manage updates.
Innovation Solution
An information processing apparatus with a model management unit, data management unit, detection unit, setting unit, and notification unit that tracks the history and updates of learning models and data, allowing users to set notification conditions and receive updates on relevant changes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If learning models and data are released by general users, then the quantity and variety of available models increase, but it becomes difficult for users to grasp the status and suitability of these models
Solution Approach 1:
The patent implements a notification system that provides feedback to users when learning models or data they have interest in are updated. The notification unit sends notifications to user terminals when detection unit detects updates to registered models or data, enabling users to grasp status changes without manually monitoring all available models among the enormous quantity released by general users.
2Reliability
If learning models and data are updated frequently, then the quality and relevance of models improve, but operational costs increase due to database storage requirements
Solution Approach 1:
The patent extracts only the essential notification function from the complete model management system. Instead of saving all inference results and model versions in a database, the system extracts and notifies only the critical information that a model or data has been updated, significantly reducing storage requirements and operational costs while maintaining the ability to inform users of quality improvements.
3Loss of information
If all learning models and data are saved in a database for tracking updates, then complete history tracking is achieved, but device complexity and operational costs increase
Solution Approach 1:
The patent extracts only the essential notification function from the complete model management system. Instead of saving all inference results and model versions in a database, the system extracts and notifies only the critical information that a model or data has been updated, significantly reducing storage requirements and operational costs while maintaining the ability to inform users of quality improvements.
Solution Approach 2:
The patent introduces a detection unit as an intermediary between the model management system and users. This intermediary detects updates and triggers notifications without requiring a comprehensive database to store all model states, simplifying the system architecture while maintaining effective update tracking and user notification capabilities.
Data Source
AI summary
An information processing apparatus manages, in association with a first learning model, information on a second learning model used for creation of the first learning model and information on a first data set used for creation of the first learning model; and manages, in association with the first data set, information on a second data set used for creation of the first data set. The apparatus detects that a third learning model created by relearning on the second learning model has been added or that a third data set created by data manipulation on the second data set has been added; sets a notification condition for notifying that the third learning model or the third data set has been added; and performs notification if the set notification condition is satisfied.