Towards Reliable Simulation-Based Inference with Balanced Neural Ratio Estimation

Arnaud Delaunoy, Joeri Hermans, François Rozet, Antoine Wehenkel, Gilles Louppe

Keywords: simulation-based inference, likelihood-free inference, approximate Bayesian inference, neural ratio estimation

TL;DR: We introduce a variation of the neural ratio estimation (NRE) algorithm for simulation-based inference that makes it more reliable in the sense that it tends to produce more conservative approximate posteriors.

Abstract: Modern approaches for simulation-based inference build upon deep learning surrogates to enable approximate Bayesian inference with computer simulators. In practice, the estimated posteriors' computational faithfulness is, however, rarely guaranteed. For example, Hermans et al., 2021 have shown that current simulation-based inference algorithms can produce posteriors that are overconfident, hence risking false inferences. In this work, we introduce Balanced Neural Ratio Estimation (BNRE), a variation of the NRE algorithm designed to produce posterior approximations that tend to be more conservative, hence improving their reliability, while sharing the same Bayes optimal solution. We achieve this by enforcing a balancing condition that increases the quantified uncertainty in low simulation budget regimes while still converging to the exact posterior as the budget increases. We provide theoretical arguments showing that BNRE tends to produce posterior surrogates that are more conservative than NRE's. We evaluate BNRE on a wide variety of tasks and show that it produces conservative posterior surrogates on all tested benchmarks and simulation budgets. Finally, we emphasize that BNRE is straightforward to implement over NRE and does not introduce any computational overhead.

Community Implementations: [ 1 code implementation](https://www.catalyzex.com/paper/arxiv:2208.13624/code)