Online tutorial: Meet ROSA, the meaningful and efficient machine learning descriptors for materials | ARC Centre of Excellence in Exciton Science
An infographic tile promoting Exciton Science’s next online tutorial. It features a navy blue background with blue, white and purple lettering, as well as the Exciton Science logo, decorative dots and a blue arrow.

Online tutorial: Meet ROSA, the meaningful and efficient machine learning descriptors for materials

30 Mar 2022

Finding the right descriptors to describe your material is the most important step in your machine learning exercise.

While there have been many material descriptors out there in the literature, each focusing on a particular aspect of the material, the pursuit of "good" descriptors is far from over.

We introduced the ROSA (rapid one-shot ab initio) descriptors that achieve the trade-off of being meaningful, as well as efficient.

In this tutorial, we will have an overview of the descriptors business in material science.

You will learn how ROSA descriptors are calculated using python.

A range of example datasets will be provided so that we can perform the machine learning training and testing together.

The ROSA descriptors are introduced in this arxiv in full detail: https://arxiv.org/abs/2203.03392

The code for the descriptors is in the Github link (https://github.com/sheriftawfikabbas/crystalfeatures), along with other descriptor classes.

Where: Online, via Zoom

When: Wednesday 30 March at 1pm AEDT.

To register, click here.