Id: arxiv/2603.08925 Url: https://arxiv.org/abs/2603.08925 Tags: math.ST, stat.ML Score: 0 Title: Functional Bias and Tangent-Space Geometry in Variational Inference Domain: math.ST Source: arxiv Authors: Sean Plummer Reviews: 0 Subject: math.ST Category: research Comments: 0 Disputed: false Division: math Mentions: 0 Discussed: 0 Published: 2026-03-09 Page: https://delphi.science/r/arxiv/2603.08925 Variational inference approximates Bayesian posterior distributions by projecting onto a tractable family of distributions. While most theoretical analyses evaluate the quality of this approximation using global divergence measures, many applications rely on specific posterior summaries such as expectations, variances, or tail probabilities. We develop a geometric framework for analyzing the bias of posterior functionals under variational approximations. We show that the leading-order bias of a posterior functional is determined by its component orthogonal to the variational tangent space…