Thread Content
Topic 1: Computational Chemistry and AI-Driven Techniques for Predicting and Screening MOF Properties】 The content covers: DFT structure optimization, molecular simulation (RASPA2), and data generation through high-throughput computing； Construct prediction models using traditional machine learning methods (XGBoost, SVM, RF) and graph neural networks (CGCNN/MEGNet) ； The structure-activity relationship is elucidated by combining explainable AI (SHAP, SISSO). The course emphasizes interdisciplinary integration and hands-on practice throughout the entire process. Through case studies and literature-based reproductions, it aims to develop comprehensive research skills ranging from data cleaning and simulation calculations to machine learning modeling and virtual screening. 【Topic 2: Artificial Intelligence and Data-Driven Methods to Accelerate the Design and Application of Metal Materials】 It systematically teaches the end-to-end approach to data-driven material design. The content covers Python for data analysis, descriptor engineering (Matminer), classical and ensemble machine learning (random forests, XGBoost), as well as advanced strategies such as active learning and multi-objective optimization. Using superalloys and titanium alloys as case studies, the course emphasizes the integration of coding practice with theory, covering the entire research process from feature extraction and model construction to hyperparameter tuning, as well as physical information/gray-box models and explainable AI (SHAP). It aims to develop the research and engineering skills needed to address practical challenges in material development. 【Topic 3: AI Applications in Polymers and Composite Materials】 Covers AI-driven design and performance prediction of polymer materials. From feature engineering to machine learning and deep learning, practical exercises focus on stress-strain prediction, generative design, and high-throughput screening, while employing tools such as SHAP for explanatory analysis to drive reverse material design. Detailed review: https://mp.weixin.qq.com/s/rsSgrmwHYLbU8pehPjHeYw Account of the author: Yan Youxin er