Riemannian Variational Flow Matching for Material and Protein Design
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866911507296550912 |
|---|---|
| author | Zaghen, Olga Eijkelboom, Floor Pouplin, Alison Liu, Cong Welling, Max van de Meent, Jan-Willem Bekkers, Erik J. |
| author_facet | Zaghen, Olga Eijkelboom, Floor Pouplin, Alison Liu, Cong Welling, Max van de Meent, Jan-Willem Bekkers, Erik J. |
| contents | We present Riemannian Gaussian Variational Flow Matching (RG-VFM), a geometric extension of Variational Flow Matching (VFM) for generative modeling on manifolds. Motivated by the benefits of VFM, we derive a variational flow matching objective for manifolds with closed-form geodesics based on Riemannian Gaussian distributions. Crucially, in Euclidean space, predicting endpoints (VFM), velocities (FM), or noise (diffusion) is largely equivalent due to affine interpolations. However, on curved manifolds this equivalence breaks down. We formally analyze the relationship between our model and Riemannian Flow Matching (RFM), revealing that the RFM objective lacks a curvature-dependent penalty -- encoded via Jacobi fields -- that is naturally present in RG-VFM. Based on this relationship, we hypothesize that endpoint prediction provides a stronger learning signal by directly minimizing geodesic distances. Experiments on synthetic spherical and hyperbolic benchmarks, as well as real-world tasks in material and protein generation, demonstrate that RG-VFM more effectively captures manifold structure and improves downstream performance over Euclidean and velocity-based baselines. Code available at https://github.com/olgatticus/rg-vfm. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2502_12981 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Riemannian Variational Flow Matching for Material and Protein Design Zaghen, Olga Eijkelboom, Floor Pouplin, Alison Liu, Cong Welling, Max van de Meent, Jan-Willem Bekkers, Erik J. Machine Learning Differential Geometry We present Riemannian Gaussian Variational Flow Matching (RG-VFM), a geometric extension of Variational Flow Matching (VFM) for generative modeling on manifolds. Motivated by the benefits of VFM, we derive a variational flow matching objective for manifolds with closed-form geodesics based on Riemannian Gaussian distributions. Crucially, in Euclidean space, predicting endpoints (VFM), velocities (FM), or noise (diffusion) is largely equivalent due to affine interpolations. However, on curved manifolds this equivalence breaks down. We formally analyze the relationship between our model and Riemannian Flow Matching (RFM), revealing that the RFM objective lacks a curvature-dependent penalty -- encoded via Jacobi fields -- that is naturally present in RG-VFM. Based on this relationship, we hypothesize that endpoint prediction provides a stronger learning signal by directly minimizing geodesic distances. Experiments on synthetic spherical and hyperbolic benchmarks, as well as real-world tasks in material and protein generation, demonstrate that RG-VFM more effectively captures manifold structure and improves downstream performance over Euclidean and velocity-based baselines. Code available at https://github.com/olgatticus/rg-vfm. |
| title | Riemannian Variational Flow Matching for Material and Protein Design |
| topic | Machine Learning Differential Geometry |
| url | https://arxiv.org/abs/2502.12981 |