Accurate fast models estimate absolute evidence for Bayesian inference
ENFR: Evidence Networks with Flow-Matched References for Amortized Simulation-Based Model Inference
Machine Learning
Summary
Sometimes, scientists want to check how well a model explains their data, which is hard to do with traditional methods. The authors developed a new technique called ENFR that quickly and accurately measures a number called evidence, which helps test if a model fits well or if it needs changes. Their method improves on previous approaches by estimating absolute evidence rather than just comparisons between models. This helps build trust in computer simulations used to understand complex data.
What this means in practice
- •For data science teams: Evaluate and validate complex simulation models faster by accurately estimating model evidence for goodness-of-fit tests.
- •For machine learning engineers: Enhance simulation-based parameter inference by integrating amortized evidence estimates to detect model misspecification automatically.
Authors
Kai Lehman, Jameson Dong, Carolina Cuesta-Lazaro, Niall Jeffrey, Benjamin D. Wandelt, Francisco Villaescusa-Navarro, Sven Krippendorf, Jochen Weller
Abstract
The evidence, $\mathcal{Z} = p(x | \mathcal{M})$, of a data-generating model $\mathcal{M}$ is a powerful tool in Bayesian frameworks: it enables model inference, model misspecification tests, and goodness-of-fit assessment as an alternative to standard discrepancy statistics, e.g. $χ^2$. While versatile, the evidence is notoriously difficult to compute, even when the likelihood is available. In both likelihood-based and simulation-based inference, estimating $\mathcal{Z}$ provides an immediate solution to high-dimensional goodness-of-fit or model misspecification tests: using $\mathcal{Z}(x)$ as the test statistic induces a minimum-volume level-$α$ test. Such tests are essential to assess model fidelity, which is necessary to trust simulation-based parameter inference. Existing learned approaches largely target Bayes factors, which only measure evidence ratios. Instead, we provide accurate, fast, and amortized estimates of the absolute evidence. We present ENFR: Evidence Networks with Flow-matched References. Building on the Evidence Network framework of Jeffrey & Wandelt (2024), the proposed method first learns Bayes factors with respect to a flow-matched reference, from which the absolute evidence is recovered. We show improvement in estimation of $\mathcal{Z}$ over the flow-based baseline in all applications considered and the best performance of all density estimation methods considered in our benchmarks. We furthermore introduce and provide the proof for an architecture that jointly learns the individual evidences of a set of models $\{\mathcal{M}_i\}$. We demonstrate the performance and scaling behavior of these approaches on data from a variety of scientific applications.