Explore data science and machine learning applications through this comprehensive hands-on training series spanning over 22 hours of practical tutorials. Master Jupyter notebooks for research workflows while learning to interact with data repositories and manage datasets effectively. Train neural networks and random forests, apply unsupervised learning techniques including Principal Component Analysis, and perform Design of Experiments using machine learning methodologies. Dive deep into materials science applications with specialized modules covering impurity level prediction in semiconductors, image segmentation using U-Net convolutional neural networks for SEM images of graphene, and quantitative structure-property relationships via materials graph networks. Develop machine learning interatomic potentials, explore physics-informed machine learning approaches, and implement Bayesian optimization for materials discovery. Learn to construct parsimonious neural networks that interpret physical laws, utilize the MAST-ML toolkit for materials property prediction, and apply Gaussian process regression for surface interpolation. Gain practical experience with autonomous experimental design, uncertainty quantification in property prediction, and large language model integration for computational research. Access specialized training in MATLAB for data analysis and machine learning, explore message-passing neural networks for molecular property prediction using Chemprop, and benchmark universal machine learning force fields. No prior coding experience required, with all exercises conducted using cloud computing resources accessible through any internet browser, enabling continued research and educational use beyond the training sessions.

Syllabus

Machine Learning Framework for Impurity Level Prediction in Semiconductors
Unsupervised Clustering Methods for Image Segmentation: Application to SEM Images of Graphene
U-Net Convolutional Neural Networks for Image Segmentation: Application to SEM Images of Graphene
Constructing Accurate Quantitative Structure-Property Relationships via Materials Graph Networks
Convenient and efficient development of Machine Learning Interatomic Potentials
Hands-on Deep Learning for Materials: Convolutional Networks and Variational Autoencoders
Batch Reification Fusion Optimization (BAREFOOT) Framework
A Hands-on Introduction to Physics-informed Machine Learning
Parsimonious Neural Networks Learn Interpretable Physical Laws
Active Learning via Bayesian Optimization for Materials Discovery
Introduction to Machine Learning for Materials Science: Workflow for Predicting Materials Properties
Materials Simulation Toolkit for Machine Learning-MAST-ML: Models for Materials Property Prediction
Parsimonious Neural Networks Learn Interpretable Physical Laws
Autonomous Neutron Diffraction Experiments with ANDiE
A Machine Learning Aided Hierarchical Screening Strategy for Materials Discovery
Debugging Neural Networks
Integrating Machine Learning with a Genetic Algorithm for Materials Exploration
Data Analysis with MATLAB
Machine Learning with MATLAB
Message-Passing Neural Networks for Molecular Property Prediction Using Chemprop
Gaussian Process Regression for Surface Interpolation
Machine Learning Predicts Additive Manufacturing Part Quality: Tutorial on Support Vector Regression
Simplifying Computational Simulations: Using Large Language Models for Automated Research in MS
Benchmarking Universal Machine Learning Force Fields with CHIPS-FF
Uncertainty in Materials Science Property Prediction: The Good, The Bad, and The Uncalibrated