Uncertainty Quantification for Distribution-to-Distribution Flow Matching in Scientific Imaging
This addresses the need for trustworthy generation in scientific imaging, such as modeling cellular responses and medical image translation, by providing a method for uncertainty quantification, though it appears incremental as it builds on existing flow matching and Bayesian techniques.
The paper tackles the problem of uncertainty quantification for distribution-to-distribution generative models in scientific imaging, proposing a framework that improves reliability and accountability, with experiments showing enhanced performance on cellular imaging and brain fMRI tasks.
Distribution-to-distribution generative models support scientific imaging tasks ranging from modeling cellular perturbation responses to translating medical images across conditions. Trustworthy generation requires both reliability (generalization across labs, devices, and experimental conditions) and accountability (detecting out-of-distribution cases where predictions may be unreliable). Uncertainty quantification (UQ) based approaches serve as promising candidates for these tasks, yet UQ for distribution-to-distribution generative models remains underexplored. We present a unified UQ framework, Bayesian Stochastic Flow Matching (BSFM), that disentangles aleatoric and epistemic uncertainty. The Stochastic Flow Matching (SFM) component augments deterministic flows with a diffusion term to improve model generalization to unseen scenarios. For UQ, we develop a scalable Bayesian approach -- MCD-Antithetic -- that combines Monte Carlo Dropout with sample-efficient antithetic sampling to produce effective anomaly scores for out-of-distribution detection. Experiments on cellular imaging (BBBC021, JUMP) and brain fMRI (Theory of Mind) across diverse scenarios show that SFM improves reliability while MCD-Antithetic enhances accountability.