Flow Annealing Posterior Sampling for Function-Space Regression and Inverse Problems
The paper introduces Flow Annealing Posterior Sampling (FAPS), a function-space posterior sampling framework for stochastic-process regression and PDE inverse problems.
FAPS is the first function-space posterior sampling framework that unifies stochastic-process regression and PDE inverse problems.
Before reading this…
Applications
- →stochastic-process regression
- →PDE inverse problems
To understand this paper, make sure you know these concepts first:
- function-space flow-matching priorsfind papers →
- Langevin correctionfind papers →
Abstract
More Like ThisPrincipled regression for stochastic processes is a long-standing challenge with deep connections to scientific inverse problems. We introduce Flow Annealing Posterior Sampling (FAPS), to our knowledge the first function-space posterior sampling framework that unifies stochastic-process regression and PDE inverse problems. Built on pretrained function-space flow-matching priors, FAPS enables likelihood-guided posterior inference from sparse and noisy observations, supports variable query discretizations, and avoids explicit prior-density evaluation. Its Langevin correction uses a low-rank covariance preconditioner to exploit dominant function-space correlations across discretizations. Across Gaussian and non-Gaussian stochastic-process regression benchmarks and diverse PDE inverse problems, FAPS produces coherent posterior samples with accurate uncertainty quantification, significantly outperforming existing functional regression baselines and achieving competitive or better PDE noisy inverse performance than diffusion-based posterior samplers while reducing test-time sampling cost.