Categorical Flow Maps
This addresses the need for faster inference in generative models for domains like images and text, though it builds incrementally on existing flow-matching and distillation techniques.
The paper tackles the problem of accelerating few-step generation of categorical data by introducing Categorical Flow Maps, a flow-matching method that achieves state-of-the-art results on images, molecular graphs, and text, with strong performance in single-step generation.
We introduce Categorical Flow Maps, a flow-matching method for accelerated few-step generation of categorical data via self-distillation. Building on recent variational formulations of flow matching and the broader trend towards accelerated inference in diffusion and flow-based models, we define a flow map towards the simplex that transports probability mass toward a predicted endpoint, yielding a parametrisation that naturally constrains model predictions. Since our trajectories are continuous rather than discrete, Categorical Flow Maps can be trained with existing distillation techniques, as well as a new objective based on endpoint consistency. This continuous formulation also automatically unlocks test-time inference: we can directly reuse existing guidance and reweighting techniques in the categorical setting to steer sampling toward downstream objectives. Empirically, we achieve state-of-the-art few-step results on images, molecular graphs, and text, with strong performance even in single-step generation.