Compositional amortized inference for large-scale hierarchical Bayesian models

Abstract: Amortized Bayesian inference (ABI) has emerged as a powerful simulation-based approach for estimating complex mechanistic models, offering fast posterior sampling via generative neural networks. However, extending ABI to hierarchical models, a cornerstone of modern Bayesian analysis, remains a major challenge due to the difficulty of scaling to large numbers of parameters. In this work, we build on compositional score matching (CSM), a divide-and-conquer strategy for Bayesian updating using diffusion models. To address existing stability issues of CSM, we propose adaptive solvers coupled with a novel, error-damping compositional estimator. Our proposed method remains stable even with hundreds of thousands of data points and parameters. We validate our approach on a controlled toy example, a high-dimensional spatial autoregressive model, and a real-world advanced microscopy biological application task involving over 750,000 parameters.