Federated Flow Matching
This work addresses the need for privacy-preserving generative model training in distributed environments, offering an incremental improvement over existing federated learning approaches.
The paper tackles the problem of training generative models on decentralized data without central aggregation by introducing Federated Flow Matching (FFM), a framework that includes methods like FFM-GOT to coordinate couplings across clients, resulting in enhanced flow straightness and sample quality comparable to centralized baselines in federated settings.
Data today is decentralized, generated and stored across devices and institutions where privacy, ownership, and regulation prevent centralization. This motivates the need to train generative models directly from distributed data locally without central aggregation. In this paper, we introduce Federated Flow Matching (FFM), a framework for training flow matching models under privacy constraints. Specifically, we first examine FFM-vanilla, where each client trains locally with independent source and target couplings, preserving privacy but yielding curved flows that slow inference. We then develop FFM-LOT, which employs local optimal transport couplings to improve straightness within each client but lacks global consistency under heterogeneous data. Finally, we propose FFM-GOT, a federated strategy based on the semi-dual formulation of optimal transport, where a shared global potential function coordinates couplings across clients. Experiments on synthetic and image datasets show that FFM enables privacy-preserving training while enhancing both the flow straightness and sample quality in federated settings, with performance comparable to the centralized baseline.