Learned Model Generation With Consent Verification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for generating learned models do not adequately address the protection of personal information, as they often use sensing data without obtaining consent from the sensing objects, which is essential for ensuring the ethical handling and usage of such data.
Innovation Solution
A method and device for generating learned models that acquire use consent from sensing objects, process or delete non-consented data, and attach electronic signatures to ensure that only consented data is used, thereby ensuring ethical data handling and usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If sensing data is used without consent to generate learned models, then productivity and efficiency of model generation is improved, but personal information protection and ethical compliance deteriorate
Solution Approach 1:
The system performs preliminary actions by acquiring consent information from sensing objects before the learned model generation process begins. The consent management module collects and stores consent status in advance, ensuring that only data from consenting individuals is included in the training dataset, thus resolving the contradiction between productivity and personal information protection.
Solution Approach 2:
The patent introduces a consent management module as an intermediary between the sensing data collection and the learned model generation processes. This intermediary component manages consent information, verifies consent status, and controls which sensing data can be used for training, thereby enabling ethical compliance without significantly impacting model generation productivity.
2Object-affected harmful factors
If consent verification processes are implemented, then personal information protection is improved, but device complexity and processing time increase
Solution Approach 1:
The system creates a consent information copy or representation that can be efficiently stored and verified. Instead of managing complex original consent documents, the system uses structured consent data representations that simplify verification processes while maintaining personal information protection, thus reducing device complexity.
3Object-affected harmful factors
If sensing data is processed or deleted based on consent status, then personal information protection is improved, but data quality and model accuracy may deteriorate
Solution Approach 1:
The system extracts and removes sensing data from individuals who have not provided consent, creating a cleaned training dataset that contains only ethically permissible data. This extraction process maintains data quality within the consented subset, ensuring model accuracy is preserved while improving personal information protection.
Solution Approach 2:
The system changes the parameter of data inclusion by filtering based on consent status. By adjusting which data points are included in the training set based on consent parameters, the system maintains high model accuracy within the ethical constraints, transforming the dataset composition without degrading overall model performance.
Data Source
AI summary
Learned model providing system has a configuration including consent acquisition device that acquires use consent that an acquired image of visitor is to be used for generating a learned model from visitor, a plurality of cameras that image visitor, learned model generating device that generates the learned model by machine learning based on a captured image imaged by camera, server device that saves the learned model generated by learned model generating device, user side device that receives the learned model from server device, camera, and member database.