Sensor Design Parameter Optimization for ML Prediction Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional technologies optimize machine learning models but fail to address the optimization of feature quantities input into these models, limiting the accuracy of sensor state discrimination.
Innovation Solution
An information processing method that acquires feature quantities from sensors, predicts the state of measurement targets using machine learning models, and determines optimal design parameter modification methods to improve prediction accuracy, thereby optimizing the feature quantities and sensor design parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If machine learning model parameters are optimized using conventional methods, then model performance is improved, but feature quantity optimization is not addressed, limiting further accuracy improvements
Solution Approach 1:
The patent applies parameter changes by systematically modifying sensor design parameters (such as measurement conditions, sensor structure, or measurement method) to optimize the feature quantities extracted from sensor data. This enables the system to adapt the input features to better suit the machine learning model, thereby improving prediction accuracy beyond what model parameter optimization alone can achieve.
2Measurement precision
If sensor design parameters are modified to improve feature quantity, then prediction accuracy is enhanced, but development complexity increases
Solution Approach 1:
The patent implements feedback by using the machine learning model's prediction results to evaluate the quality of feature quantities and guide subsequent sensor design parameter modifications. This closed-loop approach ensures that design changes are made based on actual performance impact, reducing unnecessary complexity and focusing development efforts on modifications that genuinely improve prediction accuracy.
Solution Approach 2:
The patent applies preliminary action by simulating and evaluating the impact of design parameter modifications on feature quantities before actually implementing the changes. This allows the system to identify the most effective modifications in advance, reducing development complexity by avoiding trial-and-error approaches and focusing on pre-validated design changes.
3Reliability
If multiple design parameter modification methods are evaluated, then optimum method selection is improved, but computational time increases
Solution Approach 1:
The patent applies partial action by evaluating a selected subset of the most promising design parameter modification methods rather than exhaustively analyzing all possible modifications. The system prioritizes modifications based on preliminary assessments and focuses computational resources on the most likely candidates, achieving reliable optimization results with reduced computation time.
Solution Approach 2:
The patent uses preliminary action by pre-evaluating and ranking design parameter modification methods before full optimization. This preliminary assessment identifies the most promising approaches, allowing the system to focus detailed evaluation on a limited number of candidates and significantly reducing the overall computational time required for finding the optimum method.
Data Source
AI summary
An information processing device: acquires a feature quantity indicating a feature of a measurement target measured by a sensor; predicts a state of the measurement target by inputting the feature quantity into a machine learning model; acquires a plurality of design parameter modification methods to improve state prediction accuracy of the machine learning model and modify a design parameter of the sensor; determines an optimum design parameter modification method from among the plurality of design parameter modification methods based on the feature quantity and a prediction result of the state; and outputs the optimum design parameter modification method determined.