Efficient Generative Modeling beyond Memoryless Diffusion via Adjoint Schrödinger Bridge Matching
This addresses the challenge of slow and unstable generative modeling for high-dimensional data like images, offering a more efficient alternative to existing diffusion methods.
The paper tackles the problem of inefficient and noisy sampling in diffusion models by proposing Adjoint Schrödinger Bridge Matching (ASBM), which recovers optimal trajectories to produce straighter paths, resulting in improved image generation fidelity with fewer steps and enabling distillation to a one-step generator.
Diffusion models often yield highly curved trajectories and noisy score targets due to an uninformative, memoryless forward process that induces independent data-noise coupling. We propose Adjoint Schrödinger Bridge Matching (ASBM), a generative modeling framework that recovers optimal trajectories in high dimensions via two stages. First, we view the Schrödinger Bridge (SB) forward dynamic as a coupling construction problem and learn it through a data-to-energy sampling perspective that transports data to an energy-defined prior. Then, we learn the backward generative dynamic with a simple matching loss supervised by the induced optimal coupling. By operating in a non-memoryless regime, ASBM produces significantly straighter and more efficient sampling paths. Compared to prior works, ASBM scales to high-dimensional data with notably improved stability and efficiency. Extensive experiments on image generation show that ASBM improves fidelity with fewer sampling steps. We further showcase the effectiveness of our optimal trajectory via distillation to a one-step generator.