Based on this formulation of flow matching as variational inference, CatFlow is developed, a flow matching method for categorical data that is easy to implement, computationally efficient, and achieves strong results on graph generation tasks.
This work develops this perspective theoretically, showing that quotient couplings can be lifted to aligned representatives without additional cost and that symmetrization yields equivariant flow-matching minimizers, including for categorical endpoint prediction.
Two lightweight, training-free algorithms are proposed, CAT-OV and CAT-OT that adapt step-sizes at inference time based on a novel connection between Flow Matching sampling and gradient flow that outperform existing step-size heuristics in image quality metrics across four text- to-image Flow Matching models.
Qinchan Li, Pedro Cisneros-Velarde, Ke-Ru Fu et al.· 0 citations
We propose an instantiation of flow matching that relies on a time-independent velocity field (an \emph{autonomous flow}) to exactly map between two distributions, so long as the target is singular, i.e.\ supported on a lower-dimensional data manifold. We also show that the one-step generative map associated with this...
Cheuk-Kit Lee, Florentin Coeurdoux, Yuyuan Chen et al.· 4 citations
This work proposes to amortize the graph matching (node alignment) problem and showcases the efficiency of this approach on toy and real world SGP problems of increasing complexity including a novel Mass-spectra to Scaffold task that is introduced.
F. Méndez, Paul Krzakala, Gabriel Melo et al.· 0 citations
This work introduces structured dimension-matched variational transdimensional inference (SM-VTI) for finite enumerable model spaces and develops a rooted construction graph that expresses a model as a sequence of local stop/child decisions.
Modern explicit-time generative models, such as Flow Matching [Lipman et al., 2023] and Rectified Flow [Liu et al., 2023], are typically derived top-down via Optimal Transport and the continuity equation. This standard Eulerian approach focuses on the macroscopic transport of probability mass. In this paper, we present...
P. Milanfar· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
