Category: Machine learning

Analysis software that determines 2D material structures from quantum-beam diffraction data (predecessor of ODAT-SE)

• Determine structures of surfaces and 2D materials from diffraction data

Framework for configurational sampling of multi-component disordered systems, such as metal and oxide alloys, using machine-learning models that reproduce first-principles energies

• Run configurational sampling with a model trained on first-principles energies

Optimization and inverse-problem framework for experimental data analysis (formerly 2DMAT)

• Determine structures from experimental data such as total-reflection high-energy positron diffraction
• Solve inverse problems with Bayesian optimization, replica exchange and other methods

Fast Bayesian optimization library for physics and materials science

• Speed up searches over experimental or computational candidates with Bayesian optimization

Module for using AENET (Atom-centered Neural Network) as a solver in ODAT-SE

• Optimize structural parameters using machine-learning potentials

Tool that combines HΦ and PHYSBO to determine quantum lattice model parameters from magnetization curves by Bayesian optimization

• Find Hamiltonian parameters that reproduce a target magnetization curve

Benchmark suite comparing numerical solvers for the 1D radial heat transport equation in fusion plasmas, with machine-learning solver selection

• Compare finite-volume, finite-difference, spectral and other solvers for a nonlinear heat transport equation
• Select the best solver for given parameters with machine learning

[Tutorial] Tutorial combining abICS with universal machine-learning potentials

[Related library] Cython-accelerated core modules of PHYSBO (not needed from PHYSBO v3.1.0)