Conditional Flow Matching: Simulation-Free Dynamic Optimal Transport

02/01/2023
by   Alexander Tong, et al.
0

Continuous normalizing flows (CNFs) are an attractive generative modeling technique, but they have thus far been held back by limitations in their simulation-based maximum likelihood training. In this paper, we introduce a new technique called conditional flow matching (CFM), a simulation-free training objective for CNFs. CFM features a stable regression objective like that used to train the stochastic flow in diffusion models but enjoys the efficient inference of deterministic flow models. In contrast to both diffusion models and prior CNF training algorithms, our CFM objective does not require the source distribution to be Gaussian or require evaluation of its density. Based on this new objective, we also introduce optimal transport CFM (OT-CFM), which creates simpler flows that are more stable to train and lead to faster inference, as evaluated in our experiments. Training CNFs with CFM improves results on a variety of conditional and unconditional generation tasks such as inferring single cell dynamics, unsupervised image translation, and Schrödinger bridge inference. Code is available at https://github.com/atong01/conditional-flow-matching .

READ FULL TEXT

page 20

page 24

research
10/06/2022

Flow Matching for Generative Modeling

We introduce a new paradigm for generative modeling built on Continuous ...
research
07/07/2023

Simulation-free Schrödinger bridges via score and flow matching

We present simulation-free score and flow matching ([SF]^2M), a simulati...
research
05/26/2023

Flow Matching for Scalable Simulation-Based Inference

Neural posterior estimation methods based on discrete normalizing flows ...
research
03/29/2023

Diffusion Schrödinger Bridge Matching

Solving transport problems, i.e. finding a map transporting one given di...
research
03/03/2020

Regression via Implicit Models and Optimal Transport Cost Minimization

This paper addresses the classic problem of regression, which involves t...
research
05/08/2023

CURTAINs Flows For Flows: Constructing Unobserved Regions with Maximum Likelihood Estimation

Model independent techniques for constructing background data templates ...
research
06/10/2020

Learning normalizing flows from Entropy-Kantorovich potentials

We approach the problem of learning continuous normalizing flows from a ...

Please sign up or login with your details

Forgot password? Click here to reset