Sequential Neural Likelihood: Fast Likelihood-free Inference with Autoregressive Flows
This addresses the problem of efficient likelihood-free inference for researchers in statistics and machine learning, representing a novel method for a known bottleneck rather than an incremental improvement.
The paper tackles Bayesian inference in simulator models with intractable likelihoods by proposing Sequential Neural Likelihood (SNL), which trains an autoregressive flow on simulated data to model the likelihood in high posterior density regions. The result is a method that is more robust, accurate, and requires less tuning than related neural-based approaches, with a sequential training procedure reducing simulation cost by orders of magnitude.
We present Sequential Neural Likelihood (SNL), a new method for Bayesian inference in simulator models, where the likelihood is intractable but simulating data from the model is possible. SNL trains an autoregressive flow on simulated data in order to learn a model of the likelihood in the region of high posterior density. A sequential training procedure guides simulations and reduces simulation cost by orders of magnitude. We show that SNL is more robust, more accurate and requires less tuning than related neural-based methods, and we discuss diagnostics for assessing calibration, convergence and goodness-of-fit.