Tilted Schrodinger Bridge Matching

arXiv:2609.34642v1 Announce Type: new Abstract: Schrodinger bridges provide an entropy-regularized framework and a principled solution for unpaired domain translation. In practice, a pretrained bridge may need to be adapted to human preferences or physical constraints through a reward a problem closely related to reward tilting in diffusion models but underexplored for Schrodinger bridges. We introduce Tilted Schrodinger Bridge Matching (TSBM), a post-training method for fine-tuning a learned bridge P between source p_0 and target p_1 toward a reward-tilted target p_1^rpropto p_1e^r, while preserving source p_0. We formulate this adaptation as alternating optimization initialized from P, provide theoretical justification, and derive a practical algorithm based on Adjoint Matching. We evaluate TSBM on unpaired image-to-image translation targeting digit properties in MNIST and facial attributes in CelebA.

Source: arXiv cs.LG | 2026-09-29