Home Appliance Defect Prediction Using Weighted Dual AI Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing electronic devices lack a method to predict manufacturing defects that may not be apparent during production but can occur during user operation, leading to unexpected defects and reduced product quality.
Innovation Solution
An electronic device utilizing a combination of supervised and unsupervised learning network models to predict potential defects by applying weights to prediction information from both models, based on measurement data from manufacturing and user operation stages, allowing for dynamic adjustment of prediction accuracy over the product lifecycle.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If only test and measurement values from manufacturing steps are used to determine product quality, then manufacturing efficiency is maintained, but defects may occur during user operation that were not detected
Solution Approach 1:
The patent applies preliminary action by collecting and analyzing operational data from users during the use phase, before defects actually manifest. The system continuously gathers measurement information from multiple products and uses machine learning models to predict potential defects before they occur, enabling preventive quality control rather than reactive defect detection
Solution Approach 2:
The patent implements feedback by creating a closed-loop system where operational data from users flows back to the manufacturing system. The machine learning models analyze this feedback data to identify patterns and predict defects, which then inform adjustments to manufacturing processes and quality control criteria, continuously improving product quality over time
2Measurement precision
If multiple learning network models are used to predict defects, then prediction accuracy is improved, but system complexity increases
Solution Approach 1:
The patent merges multiple learning network models (supervised learning model and unsupervised learning model) into a unified defect prediction system. The supervised learning model analyzes labeled defect data to identify known defect patterns, while the unsupervised learning model detects anomalies in unlabeled operational data. By combining the outputs of these models, the system achieves higher prediction accuracy than any single model could provide alone
Solution Approach 2:
The patent segments the defect prediction task into two distinct analytical approaches: supervised learning for known defect patterns and unsupervised learning for anomaly detection. This segmentation allows each model to specialize in its strength while working together through a weighted combination mechanism, managing complexity through functional decomposition
Data Source
AI summary
An electronic device and method for predicting whether a manufactured product will exhibit a potential defect by providing measurement information of a home appliance as input to a first learning network model and a second learning network model trained to predict whether the home appliance will exhibit a potential defect, applying a first weight to first prediction information output from the first learning network model and a second weight to second prediction information output from the second learning network model, identifying a probability that the home appliance will exhibit the potential defect based on weighted first prediction information of the first prediction information to which the first weight is applied and second prediction information of the second prediction information to which the second weight is applied. The first learning network model is a supervised learning network model and the second learning network model is an unsupervised learning network model.