Tutorials

Basics

We provide several tutorials that will help you get acquainted with DScribe and descriptors in general. If you do not have much experience with descriptors, or you simply want to refresh your memory, the tutorial on Basic concepts is a great place to start.

• Basic concepts
• Derivatives
• Sparse output

Descriptors

We have provided a basic tutorial for each descriptor included in DScribe. These tutorials briefly introduce the descriptor and demonstrate their basic call signature. We have also included several examples that should cover many of the use cases.

• Coulomb Matrix
• Sine matrix
• Ewald sum matrix
• Atom-centered Symmetry Functions
• Smooth Overlap of Atomic Positions
• Many-body Tensor Representation
• Local Many-body Tensor Representation
• Valle-Oganov descriptor

Machine Learning

In the machine learning-section you can find few real-life examples that demonstrate how DScribe can be used in machine learning applications.

• Supervised Learning: Training an ML Force-field
• Unsupervised Learning: Clustering

Visualization

Mapping the descriptors into a visual representation is useful in analyzing complex systems. These examples demonstrate how to achieve this with DScribe.

• Chemical Environment Visualization with Local Descriptors

Similarity Analysis

Measuring the similarity of structures is key to several machine learning applications — especially kernel-based methods – but these techniques are also very useful in other applications. We provide few basic tutorials on how the similarity of structures can be quantified using DScribe.

• Building similarity kernels from local environments