Medical Information Processing Device for Trained Model Verification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Machine learning models in the medical industry often deteriorate when updated, failing to match user needs and operations, leading to inaccuracies in medical data analysis.
Innovation Solution
A medical information processing device that acquires verification data sets, including medical data, result data, and true/false data, to identify a target verification data set suitable for evaluating the performance of a trained model, allowing for appropriate updates based on the relationship between the first trained model and verification data sets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a trained model is updated through additional learning to improve accuracy, then the model's general accuracy may improve, but the model may deteriorate in matching user needs and operations
Solution Approach 1:
The patent applies preliminary action by selecting and preparing verification data sets before model updating. The system identifies appropriate verification data sets that reflect user needs and operations in advance, then uses these pre-selected data sets to evaluate whether model updates will maintain alignment with user requirements, preventing deterioration before it occurs
Solution Approach 2:
The patent implements feedback by using verification data sets to evaluate model performance before and after updates. The system compares model outputs against verification data sets containing user feedback information, allowing it to determine whether updates improve or deteriorate alignment with user needs, and to reject updates that would cause deterioration
2Measurement precision
If verification data sets are comprehensively collected and evaluated, then model updating accuracy improves, but processing time and complexity increase
Solution Approach 1:
The patent applies taking out by extracting only the necessary verification data sets from the overall data collection. The system identifies and selects specific verification data sets that are most relevant to evaluating model performance for particular update scenarios, rather than comprehensively processing all available data, thus reducing processing time while maintaining evaluation accuracy
Solution Approach 2:
The patent applies partial action by using a selected subset of verification data sets rather than the complete data set. The system determines that processing a carefully selected portion of verification data sets is sufficient to accurately evaluate model updates, avoiding the excessive processing time that would result from comprehensively evaluating all available data
Data Source
AI summary
A medical information processing device of an embodiment includes processing circuitry. The processing circuitry acquires a plurality of verification data sets including medical data, result data obtained by inputting the medical data to a trained model, and true/false data regarding the result data. The processing circuit identifies a target verification data set suitable for evaluating a performance required for a trained model that outputs the result data in response to input of the medical data among the plurality of verification data sets on the basis of the relationship between a first trained model that outputs the result data in response to input of the medical data and the verification data sets.