We introduce the Approximated Optimal Transport (AOT) technique, a novel
training scheme for diffusion-based generative models. Our approach aims to
approximate and integrate optimal transport into the training process,
significantly enhancing the ability of diffusion models to estimate the
denoiser outputs accurately. This improvement leads to ODE trajectories of
diffusion models with lower curvature and reduced truncation errors during
sampling. We achieve superior image quality and reduced sampling steps by
employing AOT in training. Specifically, we achieve FID scores of 1.88 with
just 27 NFEs and 1.73 with 29 NFEs in unconditional and conditional
generations, respectively. Furthermore, when applying AOT to train the
discriminator for guidance, we establish new state-of-the-art FID scores of
1.68 and 1.58 for unconditional and conditional generations, respectively, each
with 29 NFEs. This outcome demonstrates the effectiveness of AOT in enhancing
the performance of diffusion models.

我们介绍了逼近最优输运（Approximated Optimal Transport，AOT）技术，一种用于扩散式生成模型的新型训练方案。我们的方法旨在逼近并将最优输运集成到训练过程中，大大提高了扩散模型准确估计去噪器输出的能力。通过在训练中使用 AOT，我们实现了优越的图像质量和减少的采样步骤。特别是在无条件和有条件生成中，我们分别仅使用 27 个 NFE（Numerical Fourier Expansion）和 29 个 NFE，实现了 1.88 和 1.73 的 FID（Fréchet Inception Distance）分数。此外，当将 AOT 应用于指导鉴别器的训练时，在无条件和有条件生成方面，我们分别取得了 1.68 和 1.58 的新领先 FID 分数，每个模型使用 29 个 NFE。这一结果证明了 AOT 在提升扩散模型性能方面的有效性。