Learning Model Management for Data Contributor Traceability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods fail to accurately identify and compensate users who contribute data in the creation of learning models, particularly in incremental learning scenarios, as they only track the user involved in the model itself and not the data used.
Innovation Solution
An information processing apparatus that manages learning models and data by associating identification information, enabling traceability of both the initial model and data sets used in incremental learning, allowing for proper compensation determination based on the contributions of both.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If only identification information for the user who created the model is associated with the learning model, then the model creator can be identified, but the data contributors cannot be identified and appropriate compensation cannot be set
Solution Approach 1:
The identification system is segmented into multiple levels: model identification information, data set identification information, and initial data set identification information. This segmentation allows tracking of different contributors (model creators and data contributors) separately, resolving the contradiction between identifying the model creator and tracking data contributors.
Solution Approach 2:
The patent implements a nested structure where data set identification information is embedded within the learning model information, and initial data set identification information is embedded within the data set identification information. This nested approach enables comprehensive traceability of all contributors while maintaining organized information structure.
2Reliability
If comprehensive information about all contributors is stored, then fair compensation can be determined, but the system complexity increases
Solution Approach 1:
The information management system is divided into modular components: model management information, data management information, and contribution determination information. Each module handles specific aspects of contributor tracking, reducing overall system complexity while ensuring comprehensive coverage for fair compensation.
Solution Approach 2:
The identification information structure is designed to be universal and reusable across different learning models and data sets. The same framework can track multiple contributors across multiple models without requiring separate systems, simplifying information management while ensuring comprehensive tracking.
Data Source
AI summary
An information processing apparatus for managing a plurality of learning models, comprises: a model management unit that manages, for each of the plurality of learning models, first information for identifying a data set used to learn the learning model, and second information for identifying an initial model used to learn the learning model; a data management unit that manages, for each of a plurality of data sets identified by the first information of each of the plurality of learning models, third information for identifying an initial data set used to create the data set; and a determination unit that determines, based on the first information, the second information, and the third information, at least one learning model and at least one data set used to learn a learning model of interest included in the plurality of learning models.