[Book] Materials Informatics (No. 2365)

[Available for preview] - High-speed exploration of new material candidates, prediction of compositions and structures that meet desired properties, optimization of formulation conditions.

Book Title: Development of Polymer Materials through Materials Informatics －－－－－－－－－－－－－－－－－－－－－ ■ Key Points of This Book ● Selection of Descriptors and Features! Towards Ensuring Data Quantity and Quality! - Collection of successful and unsuccessful data - Organizing data from internal and external sources such as experimental data, commercial products, and literature information - Preprocessing of noisy image data - Integration of numerical data and language data - Unit conversion, organization of measurement conditions, and documentation by experimenters - Supplementing training data with virtual data - Utilizing transfer learning and enhancing data with limited samples - Points for strengthening and expanding descriptors ● Improving Prediction Accuracy of Learning Models! - Balancing interpretability and reliability - Measures to counteract reduced interpretability due to large data volumes - Optimizing ambiguous variables through dimensionality reduction - Preprocessing explanatory variables and selecting models - Tips for incorporating scale differences into models - Ensuring reproducible data - Building multimodal AI that handles different data types such as molecules, images, and spectra - Adjusting target variables through innovative experimental and evaluation methods ● Ensuring Contradictory Physical Properties through MI, with an Eye on Scale-Up for Material Development! - Securing desired physical properties through inverse analysis

Technical and Reference Books