8 papers
Nonlinear Assimilation via Score-based Sequential Langevin Sampling
Zhao Ding, Chenguang Duan, Yuling Jiao +3
This paper introduces score-based sequential Langevin sampling (SSLS), a novel approach to nonlinear data assimilation within a recursive Bayesian filtering framework. The proposed…
Sampling via Stochastic Interpolants by Langevin-based Velocity and Initialization Estimation in Flow ODEs
Chenguang Duan, Yuling Jiao, Gabriele Steidl +3
We propose a novel method for sampling from unnormalized Boltzmann densities based on a probability flow ordinary differential equation (ODE) derived from linear stochastic interpo…
Inference-Time Alignment for Diffusion Models via Variationally Stable Doob's Matching
Jinyuan Chang, Chenguang Duan, Yuling Jiao +2
Inference-time alignment for diffusion models aims to adapt a pre-trained reference diffusion model toward a target distribution without retraining the reference score network, the…
Preconditioning and Numerical Stability in Neural Network Training for Parametric PDEs
Markus Bachmayr, Wolfgang Dahmen, Chenguang Duan +1
In the context of training neural network-based approximations of solutions of parameter-dependent PDEs, we investigate the effect of preconditioning via well-conditioned frame rep…
Characteristic Learning for Provable One Step Generation
Zhao Ding, Chenguang Duan, Yuling Jiao +3
We propose the characteristic generator, an one-step generative model that combines the sampling efficiency of generative adversarial networks (GANs) with the training stability of…
Provable Diffusion Posterior Sampling for Bayesian Inversion
Jinyuan Chang, Chenguang Duan, Yuling Jiao +3
We propose a novel diffusion-based posterior sampling method within a plug-and-play framework. Our approach constructs a probability transport from an easy-to-sample distribution t…