Diagnosis Support Apparatus Dynamic Weight Integration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional diagnosis support apparatuses face difficulties in smoothly operating when the data collection conditions change, leading to unstable inference results and increased user burden due to the need for manual integration of inference units.
Innovation Solution
The apparatus employs first, second, and third inference means, each using data collected under different conditions, with integration means that use weights to combine their results, allowing for seamless operation across varying conditions without user intervention.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the inference unit is sequentially updated using newly collected data, then the inference unit adapts to new data trends, but the inference result becomes unstable and the inference unit updates all the time
Solution Approach 1:
The patent implements dynamic weight adjustment where the weight of the inference unit is automatically modified based on the degree of change in data trends. When data distribution changes significantly, the weight is reduced to allow adaptation; when data distribution remains stable, the weight is maintained to preserve inference stability. This dynamic balancing mechanism resolves the contradiction between adaptability and stability.
2Adaptability or versatility
If the user sets a method of integrating the results of the inference units, then the integration can be customized, but the user burden increases
Solution Approach 1:
The patent implements automatic weight adjustment mechanisms that perform integration without requiring user settings. The system autonomously determines the appropriate weight for each inference unit based on data distribution analysis and change detection, eliminating the need for users to configure integration methods while still achieving adaptive integration results.
3Adaptability or versatility
If the diagnosis support apparatus is operated under different conditions from those used for learning data collection, then the apparatus can be deployed more widely, but correct inference becomes difficult
Solution Approach 1:
The patent monitors changes in data distribution parameters when operating under different conditions and dynamically adjusts the weight of the inference unit based on these parameter changes. When significant deviations from learning conditions are detected, the system reduces reliance on the original inference unit and increases weight for retrained units, thereby maintaining inference accuracy across varying operational conditions.
Data Source
AI summary
A diagnosis support apparatus comprising: first inference means for performing inference concerning diagnosis using data collected under a first condition; second inference means for performing inference concerning the diagnosis using the data collected under the first condition and data collected under a second condition different from the first condition; third inference means for performing inference concerning the diagnosis using the data collected under the second condition; and integration means for integrating inference results of the first inference means, the second inference means, and the third inference means using a first weight, a second weight, and a third weight respectively corresponding to the first inference means, the second inference means, and the third inference means.