A Trust Crisis In Simulation-Based Inference? Your Posterior Approximations Can Be Unfaithful

TL;DR

This paper demonstrates that current Bayesian simulation-based inference algorithms often produce overconfident and unreliable posterior approximations, risking their scientific validity and urging the development of more conservative methods.

Contribution

The study provides extensive empirical evidence of unfaithfulness in existing algorithms and suggests ensembling as a promising approach to improve reliability.

Findings

All benchmarked algorithms can produce overconfident posteriors

Ensembling posterior surrogates improves approximation reliability

Addressing unfaithfulness is crucial for scientific applications

Abstract

We present extensive empirical evidence showing that current Bayesian simulation-based inference algorithms can produce computationally unfaithful posterior approximations. Our results show that all benchmarked algorithms -- (Sequential) Neural Posterior Estimation, (Sequential) Neural Ratio Estimation, Sequential Neural Likelihood and variants of Approximate Bayesian Computation -- can yield overconfident posterior approximations, which makes them unreliable for scientific use cases and falsificationist inquiry. Failing to address this issue may reduce the range of applicability of simulation-based inference. For this reason, we argue that research efforts should be made towards theoretical and methodological developments of conservative approximate inference algorithms and present research directions towards this objective. In this regard, we show empirical evidence that ensembling posterior surrogates provides more reliable approximations and mitigates the issue.