NICE

NICE (N-body Iteratively Contracted Equivariants) is a set of tools designed for the calculation of invariant and covariant atomic structure representations. It allows to automatically select the most informative combinations of high-order spectrum elements and perform their efficient computation using recurrence relations.

Though being designed specifically for atomistic machine learning, NICE, in principle, can be applied to other machine learning tasks towards producing invariant or covariant output that involve signals within a spherical cutoff, either on the sphere surface (2D manifold) or within the enclosed area (3D manifold).

Installation

1. Install librascal

2. git clone or download archive with nice and unpack

3. cd to root nice directory and run pip3 install .

Theory in a nutshell

• Theory in a nutshell

Tutorials

• Constructing machine learning potential
• Calculating covariants
• Getting insights about the model
• Constructor or non standard sequence
• Sequential fitting
• Custom regressors into purifiers

Examples

• Real-Life Examples of Nice

Reference guide

• Blocks
  • NICE
  • Compressors
  • Expansioners
  • Purifiers
  • Miscellaneous
• Utilities

Acknowledgements

• Acknowledgements