Abstract:Recently, deep generative models have been used for posterior inference in inverse problems, including high-stakes applications in medical imaging and scientific discovery, where the uncertainty of a prediction can matter as much as the prediction itself. However, posterior uncertainty is difficult to interpret because it can mix ambiguity inherent to the forward operator with uncertainty propagated through inference. We introduce a structural decomposition of posterior uncertainty that isolates intrinsic ambiguity. A cascade formulation makes this ambiguity accessible for calibration analysis, enabling qualitative diagnostics and simulation-based calibration tests that reveal failure modes that remain hidden when models are selected by reconstruction quality alone. We first validate the approach on a Gaussian example with analytical posterior structure, then illustrate the decomposition on accelerated magnetic resonance imaging (MRI), and finally apply the calibration diagnostics to electroencephalography (EEG) source imaging.
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2605.15050 [cs.LG] |
| (or arXiv:2605.15050v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2605.15050 arXiv-issued DOI via DataCite (pending registration) |
From: Yuxin Guo [view email]
[v1]
Thu, 14 May 2026 16:45:19 UTC (6,225 KB)
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。