Reproduce “Explainable Artificial Intelligence for Manufacturing Cost Estimation and Machining Feature Visualization” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Soyoung Yoo, Namwoo Kang
Researchers developed an explainable deep learning framework that estimates manufacturing costs for 3D CAD models with a Mean Absolute Percentage Error (MAPE) of 8.76% and visualizes cost-driving machining features using 3D Grad-CAM. The approach implicitly learns both specific machining features and their difficulty levels, offering actionable insights for design optimization.
Key insights
- Integrating non-geometric data (material type, part volume) with 3D CAD representations significantly enhances the accuracy of manufacturing cost prediction.
- Deep learning models can implicitly learn to detect CNC machining features and differentiate varying levels of machining difficulty solely through the task of cost prediction, without explicit feature recognition training.
- 3D Grad-CAM provides actionable visual explanations of cost drivers, enabling designers to identify and modify cost-critical design elements during early product development.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.