Apax: A Flexible and Performant Framework for the Development of Machine-Learned Interatomic Potentials

  • 0Institute for Theoretical Chemistry, University of Stuttgart, Pfaffenwaldring 55, 70569 Stuttgart, Germany.

|

|

Summary

This summary is machine-generated.

Atomistic learned potentials in JAX (apax) is a new open-source software package for machine-learned interatomic potentials. It offers efficient training and inference, accelerating materials discovery through active learning and improved model accuracy.

Area Of Science

  • Computational materials science
  • Machine learning for scientific computing

Background

  • Machine-learned interatomic potentials (MLIPs) are crucial for simulating materials at the atomic scale.
  • Existing MLIP frameworks can be computationally intensive and lack flexibility for advanced applications like active learning.

Purpose Of The Study

  • Introduce Atomistic learned potentials in JAX (apax), an open-source software package for MLIPs.
  • Demonstrate apax's capabilities in accelerating materials simulations and active learning workflows.

Main Methods

  • Developed apax leveraging the JAX framework for GPU acceleration and flexible model design.
  • Implemented features for active learning: kernel-based data selection, uncertainty estimation, and enhanced sampling.
  • Trained models on ionic liquid and Li3PO4 datasets, comparing performance against existing methods.

Main Results

  • Continuously learning models in apax reduced training times by up to 85% for active learning.
  • A Gaussian Moment Neural Network in apax achieved higher accuracy and 10x faster inference than Allegro.
  • Demonstrated good scalability in data-parallel training and modularity for complex model ensembles.

Conclusions

  • apax provides a flexible, efficient, and user-friendly platform for developing and deploying MLIPs.
  • The software facilitates advanced applications like active learning, leading to faster and more accurate materials simulations.
  • apax enables the development of novel MLIP models and accelerates the pace of materials discovery.

Related Concept Videos

Atomic Orbitals 02:44

35.6K

An atomic orbital represents the three-dimensional regions in an atom where an electron has the highest probability to reside. The radial distribution function indicates the total probability of finding an electron within the thin shell at a distance r from the nucleus. The atomic orbitals have distinct shapes which are determined by l, the angular momentum quantum number. The orbitals are often drawn with a boundary surface, enclosing densest regions of the cloud. 

The angular momentum...

Predicting Molecular Geometry 02:27

36.1K

VSEPR Theory for Determination of Electron Pair Geometries

The following procedure uses VSEPR theory to determine the electron pair geometries and the molecular structures:


Write the Lewis structure of the molecule or polyatomic ion.
Count the number of electron groups (lone pairs and bonds) around the central atom. A single, double, or triple bond counts as one region of electron density.
Identify the electron-pair geometry based on the number of electron groups: linear, trigonal planar,...

Electronic Structure of Atoms 02:28

24.3K


An atom comprises protons and neutrons, which are contained inside the dense, central core called the nucleus, with electrons present around the nucleus. Taking into account the wave–particle duality of electrons and the uncertainty in position around the nucleus, quantum mechanics provides a more accurate model for the atomic structure. It describes atomic orbitals as the regions around the nucleus where electrons of discrete energy exist, characterized by four quantum...

Thermodynamic Potentials 01:26

973

Thermodynamic potentials are state functions that are extremely useful in analyzing a thermodynamic system. They have dimensions of energy. The four important thermodynamic potentials are internal energy, enthalpy, Helmholtz free energy, and Gibbs free energy. These thermodynamic potentials can be expressed using two of the following variables: pressure, volume, temperature, and entropy. These two variables are expressed as the rate of change of the thermodynamic potential with respect to other...

Force and Potential Energy in One Dimension 01:13

5.5K

Force can be calculated from the expression for potential energy, which is a function of position. The component of a conservative force, in a particular direction, equals the negative of the derivative of the corresponding potential energy with respect to the displacement in that direction. For regions where potential energy changes rapidly with displacement, the work done and force is maximum. Also, when force is applied along the positive coordinate axis, the potential energy decreases with...

Molecular Orbital Theory I 02:35

32.9K

Overview of Molecular Orbital Theory

Molecular orbital theory describes the distribution of electrons in molecules in the same way as the distribution of electrons in atoms is described using atomic orbitals. Quantum mechanics describes the behavior of an electron in a molecule by a wave function, Ψ, analogous to the behavior in an atom. Just like electrons around isolated atoms, electrons around atoms in molecules are limited to discrete (quantized) energies. The region of space in which...