Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Improving Industrial Injection Molding Processes with Explainable AI for Quality Classification

Georg Rottenwalter, Marcel Tilly, Victor Owolabi

TL;DR

This work tackles the interpretability barrier in industrial ML for injection-molding quality classification by applying SHAP, Grad-CAM, and LIME to an LSTM time-series model. By reducing the input features from 19 to 9 and 6 based on XAI insights, the authors show that a carefully selected subset can maintain or even improve predictive performance while enhancing interpretability and inference speed. The 9-feature model achieves the best average performance (accuracy ~91%, F1 ~92%), with notable improvements in generalization and a clear view of influential process parameters; the 6-feature model remains strong but less stable. These findings support more feasible AI-driven quality control in industry, particularly for facilities with limited sensing capabilities and a path toward efficient synthetic data generation using a smaller, informative feature set.

Abstract

Machine learning is an essential tool for optimizing industrial quality control processes. However, the complexity of machine learning models often limits their practical applicability due to a lack of interpretability. Additionally, many industrial machines lack comprehensive sensor technology, making data acquisition incomplete and challenging. Explainable Artificial Intelligence offers a solution by providing insights into model decision-making and identifying the most relevant features for classification. In this paper, we investigate the impact of feature reduction using XAI techniques on the quality classification of injection-molded parts. We apply SHAP, Grad-CAM, and LIME to analyze feature importance in a Long Short-Term Memory model trained on real production data. By reducing the original 19 input features to 9 and 6, we evaluate the trade-off between model accuracy, inference speed, and interpretability. Our results show that reducing features can improve generalization while maintaining high classification performance, with an small increase in inference speed. This approach enhances the feasibility of AI-driven quality control, particularly for industrial settings with limited sensor capabilities, and paves the way for more efficient and interpretable machine learning applications in manufacturing.

Improving Industrial Injection Molding Processes with Explainable AI for Quality Classification

TL;DR

This work tackles the interpretability barrier in industrial ML for injection-molding quality classification by applying SHAP, Grad-CAM, and LIME to an LSTM time-series model. By reducing the input features from 19 to 9 and 6 based on XAI insights, the authors show that a carefully selected subset can maintain or even improve predictive performance while enhancing interpretability and inference speed. The 9-feature model achieves the best average performance (accuracy ~91%, F1 ~92%), with notable improvements in generalization and a clear view of influential process parameters; the 6-feature model remains strong but less stable. These findings support more feasible AI-driven quality control in industry, particularly for facilities with limited sensing capabilities and a path toward efficient synthetic data generation using a smaller, informative feature set.

Abstract

Machine learning is an essential tool for optimizing industrial quality control processes. However, the complexity of machine learning models often limits their practical applicability due to a lack of interpretability. Additionally, many industrial machines lack comprehensive sensor technology, making data acquisition incomplete and challenging. Explainable Artificial Intelligence offers a solution by providing insights into model decision-making and identifying the most relevant features for classification. In this paper, we investigate the impact of feature reduction using XAI techniques on the quality classification of injection-molded parts. We apply SHAP, Grad-CAM, and LIME to analyze feature importance in a Long Short-Term Memory model trained on real production data. By reducing the original 19 input features to 9 and 6, we evaluate the trade-off between model accuracy, inference speed, and interpretability. Our results show that reducing features can improve generalization while maintaining high classification performance, with an small increase in inference speed. This approach enhances the feasibility of AI-driven quality control, particularly for industrial settings with limited sensor capabilities, and paves the way for more efficient and interpretable machine learning applications in manufacturing.

Paper Structure

This paper contains 17 sections, 2 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Explainable AI background
  4. Explainable AI in the industry
  5. AI in injection molding
  6. Methods
  7. Training Evaluation Data
  8. Parameter analysis
  9. Experimental setup
  10. Training & validation data
  11. LSTM Training
  12. XAI parameter analysis
  13. Results
  14. XAI results
  15. Influence of feature reduction on model performance
  16. ...and 2 more sections

Figures (2)

  • Figure 1: System design of overall concept.
  • Figure 2: Aggregated feature maps (Grad-CAM top, LIME middle, SHAP bottom) showing average feature relevance across time bins of the injection molding cycle.