Car Body Model Learning System Optimizing Design With Limited Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Developing a car body using neural network learning is challenging due to the difficulty in securing a large amount of data, and the goal is to acquire an optimal specification for design rather than predicting target values accurately.
Innovation Solution
A model learning system and method for car body development that acquires data around the optimal specification, sets a target area based on minimum data, and redefines the loss function within this target area to derive optimal specifications for car body design.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional neural network learning is used for car body development, then accurate prediction of target values can be achieved, but a large amount of data is required which is difficult to secure
Solution Approach 1:
The patent applies local quality by defining a target area around the optimal specification point in the parameter space. Instead of requiring large amounts of data across the entire design space, the system concentrates learning on a localized region (target area) where the optimal car body specification exists. This is achieved by setting a target area based on initial model learning results and then redefining the loss function to focus exclusively on minimizing errors within this localized region, thereby achieving high prediction accuracy with limited data.
Solution Approach 2:
The patent changes the parameter of the loss function from a global error metric to a localized error metric. By redefining the loss function to calculate errors only within the target area rather than across the entire data distribution, the system transforms the learning objective to focus on the critical region. This parameter change in the loss function definition enables the model to achieve accurate predictions for optimal specifications using significantly fewer training data points.
2Adaptability or versatility
If data is collected across the entire design space, then comprehensive coverage is achieved, but the data amount becomes unmanageably large for car body development
Solution Approach 1:
The patent extracts the critical region from the entire design space by defining a target area around the optimal specification point. Instead of utilizing data from the complete design space, the system selectively extracts and focuses on the relevant portion (target area) where optimal car body specifications are located. This extraction approach maintains design adaptability by concentrating resources on the most important region while dramatically reducing the total data quantity required.
Solution Approach 2:
The patent segments the design space into a target area and the rest of the space. By dividing the continuous parameter space and focusing learning only on the segmented target region (defined by ±2ΔNRMSE from the optimal point), the system achieves comprehensive coverage of the critical design region without needing data from the entire design space. This segmentation enables efficient use of limited data resources.
Data Source
AI summary
A model learning system and a method for car body development are disclosed. The model learning system includes a data acquisition module configured to acquire data for performing learning of a model for car body development. The model learning system also includes a model learning module configured to perform the model learning by using the acquired data. The model learning system further includes a target area setting module configured to set a target area from a result of the model learning. The model learning system further includes a loss function improvement module configured to improv a loss function based on the target area. The model learning system further still includes an optimal specification derivation module deriving an optimal specification for a car body in development by using the model learned based on the improved loss function.