Simulation-based inference using surjective sequential neural likelihood estimation

arXiv:2308.01054v2 Announce Type: replace-cross
Abstract: We present Surjective Sequential Neural Likelihood (SSNL) estimation, a novel method for simulation-based inference in models where the evaluation of the likelihood function is not tractable and only a simulator that can generate synthetic data is available. SSNL fits a dimensionality-reducing surjective normalizing flow model and uses it as a surrogate likelihood function which allows for conventional Bayesian inference using either Markov chain Monte Carlo methods or variational inference. By embedding the data in a …

abstract arxiv cs.lg data dimensionality evaluation flow function generate inference likelihood novel simulation stat.me stat.ml synthetic synthetic data tractable type