Machine Learning-Based Manufacturing Cost Prediction from 2D Engineering Drawings via Geometric Features

We present an integrated machine learning framework that transforms how manufacturing cost is estimated from 2D engineering drawings. Unlike traditional quotation workflows that require labor-intensive process planning, our approach about 200 geometric and statistical descriptors directly from 13,684 DWG drawings of automotive suspension and steering parts spanning 24 product groups. Gradient-boosted decision tree models (XGBoost, CatBoost, LightGBM) trained on these features achieve nearly 10% mean absolute percentage error across groups, demonstrating robust scalability beyond part-specific heuristics. By coupling cost prediction with explainability tools such as SHAP, the framework identifies geometric design drivers including rotated dimension maxima, arc statistics and divergence metrics, offering actionable insights for cost-aware design. This end-to-end CAD-to-cost pipeline shortens quotation lead times, ensures consistent and transparent cost assessments across part families and provides a deployable pathway toward real-time, ERP-integrated decision support in Industry 4.0 manufacturing environments.

Accelerating nanomaterials discovery with artificial intelligence at the HPC centers

Study of properties of chemicals, drugs, biomaterials and alloys requires decades of dedicated work. Often times the outcome, however, is not what is expected for practical applications. This research procedure can be inverted by the new artificial intelligence and optimization methods. Instead of studying the properties of a material and its structurally close derivatives, the chemical and structural parameter space that contains all possible derivatives of that material can be scanned in a fast and smart way at the HPC centers. As a result of this, the particular material that has the specific physical or chemical properties can be found. Here we show how Bayesian optimization, Gaussian regression and artificial neural networks can be used towards this goal. We present an example smart search on the doped graphene quantum dot parameter space.