Generative Modeling through Koopman Spectral Analysis: An Operator-Theoretic Perspective
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908818901827584 |
|---|---|
| author | Xu, Yuanchao Li, Fengyi Fujisawa, Masahiro Cheng, Xiaoyuan Marzouk, Youssef Ishikawa, Isao |
| author_facet | Xu, Yuanchao Li, Fengyi Fujisawa, Masahiro Cheng, Xiaoyuan Marzouk, Youssef Ishikawa, Isao |
| contents | We propose Koopman Spectral Wasserstein Gradient Descent (KSWGD), a particle-based generative modeling framework that learns the Langevin generator via Koopman theory and integrates it with Wasserstein gradient descent. Our key insight is that this spectral structure of the underlying distribution can be directly estimated from trajectory data via the Koopman operator, eliminating the need for explicit knowledge of the target potential. Additionally, we prove that KSWGD maintains an approximately constant dissipation rate, thereby establishing linear convergence and overcoming the vanishing-gradient phenomenon that hinders existing kernel-based particle methods. We further provide a Feynman--Kac interpretation that clarifies the method's probabilistic foundation. Experiments on compact manifolds, metastable multi-well systems, and high-dimensional stochastic partial differential equations demonstrate that KSWGD consistently outperforms baselines in both convergence speed and sample quality. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_18837 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Generative Modeling through Koopman Spectral Analysis: An Operator-Theoretic Perspective Xu, Yuanchao Li, Fengyi Fujisawa, Masahiro Cheng, Xiaoyuan Marzouk, Youssef Ishikawa, Isao Machine Learning Dynamical Systems We propose Koopman Spectral Wasserstein Gradient Descent (KSWGD), a particle-based generative modeling framework that learns the Langevin generator via Koopman theory and integrates it with Wasserstein gradient descent. Our key insight is that this spectral structure of the underlying distribution can be directly estimated from trajectory data via the Koopman operator, eliminating the need for explicit knowledge of the target potential. Additionally, we prove that KSWGD maintains an approximately constant dissipation rate, thereby establishing linear convergence and overcoming the vanishing-gradient phenomenon that hinders existing kernel-based particle methods. We further provide a Feynman--Kac interpretation that clarifies the method's probabilistic foundation. Experiments on compact manifolds, metastable multi-well systems, and high-dimensional stochastic partial differential equations demonstrate that KSWGD consistently outperforms baselines in both convergence speed and sample quality. |
| title | Generative Modeling through Koopman Spectral Analysis: An Operator-Theoretic Perspective |
| topic | Machine Learning Dynamical Systems |
| url | https://arxiv.org/abs/2512.18837 |