BridgeTwist: Twisting Schrödinger Bridges for Training-Free Conditional Sampling
Abstract
Generative models are often required to serve many downstream conditional tasks. In the training-free setting, a pretrained unconditional model is reused as a prior and combined with a task-specific likelihood to sample from the corresponding posterior. While such training-free samplers have been developed for diffusion and flow models, their counterpart for Schrödinger bridge (SB) models remains unexplored, despite the growing appeal of SBs as flexible stochastic transport models between general source and target distributions. We propose BridgeTwist, a training-free conditional sampler for pretrained SB generative models, formulated as a tempered twisted sequential Monte Carlo scheme. Its twist is a closed-form plug-in surrogate, obtained by combining the velocity and score fields exposed by SB pretraining into an endpoint predictor and evaluating the likelihood there. Since this surrogate is less reliable near the source, we temper the twist with a time-varying exponent, downweighting its influence at early steps. We further introduce a bridge-shaped inference-time diffusion schedule that concentrates the proposal noise in the interior of the path. Telescoping importance weights then cancel all intermediate twist factors, yielding an asymptotically exact sampler for the posterior. Empirically, BridgeTwist outperforms training-free baselines on class-conditional and text-prompted image generation, and is competitive with specialized inverse problem solvers on inpainting, deblurring, and super-resolution.
Related
- Likelihood-Free Schrödinger--Föllmer Sampling for Bayesian Inference with Fixed-Proposal Reuse
- Quantum-Enhanced Sampling of Schrödinger Bridges
- Total variation bounds for Schrödinger--Föllmer sampling of boundary-vanishing densities
- Conditional Decomposition Discrepancy: Separating Conditional Bias from Sampling Bias in Small-Sample Cross-Domain Cost Forecasting
- Direct Conditional Transition Sampling for Diffusion Inverse Problems
BibTeX
@misc{li2026bridgetwist,
title = {{BridgeTwist: Twisting Schrödinger Bridges for Training-Free Conditional Sampling}},
author = {Mengyu Li and Qianqian Qu and Jun S. Liu},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.18658},
eprint = {2610.18658},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.18658},
}
RIS
TY - UNPB
TI - BridgeTwist: Twisting Schrödinger Bridges for Training-Free Conditional Sampling
AU - Mengyu Li
AU - Qianqian Qu
AU - Jun S. Liu
PY - 2026
DA - 2026/10/13
PB - arXiv
DO - 10.48550/arxiv.2610.18658
UR - https://arxiv.org/abs/2610.18658
AB - Generative models are often required to serve many downstream conditional tasks. In the training-free setting, a pretrained unconditional model is reused as a prior and combined with a task-specific likelihood to sample from the corresponding posterior. While such training-free samplers have been developed for diffusion and flow models, their counterpart for Schrödinger bridge (SB) models remains unexplored, despite the growing appeal of SBs as flexible stochastic transport models between general source and target distributions. We propose BridgeTwist, a training-free conditional sampler for pretrained SB generative models, formulated as a tempered twisted sequential Monte Carlo scheme. Its twist is a closed-form plug-in surrogate, obtained by combining the velocity and score fields exposed by SB pretraining into an endpoint predictor and evaluating the likelihood there. Since this surrogate is less reliable near the source, we temper the twist with a time-varying exponent, downweighting its influence at early steps. We further introduce a bridge-shaped inference-time diffusion schedule that concentrates the proposal noise in the interior of the path. Telescoping importance weights then cancel all intermediate twist factors, yielding an asymptotically exact sampler for the posterior. Empirically, BridgeTwist outperforms training-free baselines on class-conditional and text-prompted image generation, and is competitive with specialized inverse problem solvers on inpainting, deblurring, and super-resolution.
ER -