Follow us


Type: Artificial Intelligence applications (Likelihood-free inference package)
Authors: Guo-Jian Wang
Abstract: CoLFI (Likelihood-free inference package for cosmology and the wider sciences) is a framework to estimate model parameters based on neural density estimators. It is an alternative to the traditional Markov chain Monte Carlo (MCMC) method. CoLFI models parameters directly without using Bayes's theorem. Therefore, in a narrow sense, it is independent of Bayesian theory.