Skip to content

Gaussian Process Bayesian Toolkit with Monte Carlo Sampler Integration for Heavy Ion Collisions

License

Notifications You must be signed in to change notification settings

Hendrik1704/GPBayesTools-HIC

Repository files navigation

GPBayesTools-HIC

Gaussian Process Bayesian Toolkit with Monte Carlo Sampler Integration for Heavy Ion Collisions

This toolkit implements a wrapper for Gaussian Process (GP) emulators and Monte Carlo (MC) samplers used in high-energy heavy-ion simulations.

The following wrappers for GP emulators are currently included:

  • Scikit Learn GP emulator wrapper
  • PCGP and PCSK wrapper for the GPs implemented in the surmise package of the BAND Collaboration

The following wrappers for MC sampling are included:

  • MCMC wrapper for the emcee package
  • PTMCMC (Parallel Tempering Markov Chain Monte Carlo) wrapper
    • Not recommend to use this one for larger runs. There are problems with the parallelization.
  • PTLMC from the surmise package (Parallel Tempering Langevin Monte Carlo)
  • pocoMC Preconditioned Monte Carlo method for accelerated Bayesian inference

❗ The jupyter notebooks are just meant as examples for how to use the emulators and samplers and analyze the output. Paths and data files need the proper input formats.

About

Gaussian Process Bayesian Toolkit with Monte Carlo Sampler Integration for Heavy Ion Collisions

Topics

Resources

License

Stars

Watchers

Forks

Packages

No packages published