Machine Learning (ML) and Artificial Intelligence (AI) are powerful techniques that materials scientists can use to help analyze their data, choose experiments, and discover new materials. This tutorial will introduce basic techniques for ML and AI, all from a materials science perspective. Part of the purpose of this tutorial will be explaining how many of these techniques work, dispelling myths arising from the hype from popular culture. We will also show how these tools can be used for more rigorous materials science studies, and how doing so differs from the prototypical ML and AI methods designed by computer science and social for use with largely unstructured. We show how to adapt the ML and AI methods to the particular data challenges in materials science with the goals of answering scientific inquiries.  

After the tutorial, attendees will be familiar with and have the resources to:

• Apply the basics of both supervised and unsupervised ML and AI techniques.
• Apply Gaussian Processes and Active Learning to materials science problems.
• Use Deep Learning techniques to analyze large data sets.

Tutorial Schedule

8:00 am

Introduction to Machine Learning and Artificial Intelligence

Arun Mannodi Kanakkithodi, Purdue University

9:30 am BREAK

10:00 am

Gaussian Process Regression and Active Learning

Austin McDannald, National Institute of Standards and Technology

12:00 pm BREAK

1:30 pm

Deep Learning With Neural Networks

Saaketh Desai, Sandia National Laboratories

2:30 pm DISCUSSION

3:00 pm BREAK

3:30 pm

Hands-on Machine Learning Competition With Materials Science Data

Tyler Martin, National Institute of Standards and Technology; Peter Beaucage; National Institute of Standards and Technology

Announcing the 3rd Annual MLMR Materials Informatics Competition!

Each year MLMR teams up with a different industry partner to bring you a real-world materials informatics challenge. This year’s challenge is brought to you through a collaboration with the nSOFT Consortium, financial sponsors Cell Matter and Patterns, and support from the Materials Research Society. We present a challenge in Active Learning - the machine learning field used to drive recommendation engines and autonomous systems. The challenge runs April 22 - May 6, 2024 with $1000 in prizes with multiple winners! Click here for information and registration »