Guardado en:
| Autores principales: | , , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | https://arxiv.org/abs/2602.08243 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866915785723609088 |
|---|---|
| author | Guo, Wei Zhu, Yuchen Du, Xiaochen Nam, Juno Chen, Yongxin Gómez-Bombarelli, Rafael Liu, Guan-Horng Tao, Molei Choi, Jaemoo |
| author_facet | Guo, Wei Zhu, Yuchen Du, Xiaochen Nam, Juno Chen, Yongxin Gómez-Bombarelli, Rafael Liu, Guan-Horng Tao, Molei Choi, Jaemoo |
| contents | Learning discrete neural samplers is challenging due to the lack of gradients and combinatorial complexity. While stochastic optimal control (SOC) and Schrödinger bridge (SB) provide principled solutions, efficient SOC solvers like adjoint matching (AM), which excel in continuous domains, remain unexplored for discrete spaces. We bridge this gap by revealing that the core mechanism of AM is $\mathit{state}\text{-}\mathit{space~agnostic}$, and introduce $\mathbf{discrete~ASBS}$, a unified framework that extends AM and adjoint Schrödinger bridge sampler (ASBS) to discrete spaces. Theoretically, we analyze the optimality conditions of the discrete SB problem and its connection to SOC, identifying a necessary cyclic group structure on the state space to enable this extension. Empirically, discrete ASBS achieves competitive sample quality with significant advantages in training efficiency and scalability. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_08243 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Discrete Adjoint Schrödinger Bridge Sampler Guo, Wei Zhu, Yuchen Du, Xiaochen Nam, Juno Chen, Yongxin Gómez-Bombarelli, Rafael Liu, Guan-Horng Tao, Molei Choi, Jaemoo Machine Learning Learning discrete neural samplers is challenging due to the lack of gradients and combinatorial complexity. While stochastic optimal control (SOC) and Schrödinger bridge (SB) provide principled solutions, efficient SOC solvers like adjoint matching (AM), which excel in continuous domains, remain unexplored for discrete spaces. We bridge this gap by revealing that the core mechanism of AM is $\mathit{state}\text{-}\mathit{space~agnostic}$, and introduce $\mathbf{discrete~ASBS}$, a unified framework that extends AM and adjoint Schrödinger bridge sampler (ASBS) to discrete spaces. Theoretically, we analyze the optimality conditions of the discrete SB problem and its connection to SOC, identifying a necessary cyclic group structure on the state space to enable this extension. Empirically, discrete ASBS achieves competitive sample quality with significant advantages in training efficiency and scalability. |
| title | Discrete Adjoint Schrödinger Bridge Sampler |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2602.08243 |