Transition Flow Matching learns a global transition flow rather than local velocity fields, enabling single-step generation and transfer to arbitrary future time points.
March 18, 2026
Original Paper
Transition Flow Matching
arXiv · 2603.15689
The Takeaway
This unifies several flow-based paradigms and eliminates the need for multiple integration steps, potentially making flow-matching models as efficient as GANs or one-step diffusers.
From the abstract
Mainstream flow matching methods typically focus on learning the local velocity field, which inherently requires multiple integration steps during generation. In contrast, Mean Velocity Flow models establish a relationship between the local velocity field and the global mean velocity, enabling the latter to be learned through a mathematically grounded formulation and allowing generation to be transferred to arbitrary future time points. In this work, we propose a new paradigm that directly learn