FFJORD：可伸缩可逆生成模型的自由形式连续动力学
FFJORD: Free-form Continuous Dynamics for Scalable Reversible Generative Models
摘要：
一类有前途的生成模型通过可逆神经网络将点从简单分布映射到复杂分布。这些模型的基于可能性的训练需要限制它们的架构，以允许雅可比行列式的廉价计算。或者，如果变换由常微分方程指定，则可以使用雅可比迹。在本文中，我们使用Hutchinson的迹估计给出了一个可扩展的无偏估计的对数密度。其结果是一个连续时间可逆的生成模型，无偏密度估计和一次采样，同时允许不受限制的神经网络架构。我们展示了我们的方法在高维密度估计，图像生成和变分推理，实现国家的最先进的精确似然方法与有效的采样。
英文摘要：
A promising class of generative models maps points from a simple distribution to a complex distribution through an invertible neural network. Likelihood-based training of these models requires restricting their architectures to allow cheap computation of Jacobian determinants. Alternatively, the Jacobian trace can be used if the transformation is specified by an ordinary differential equation. In this paper, we use Hutchinson's trace estimator to give a scalable unbiased estimate of the log-density. The result is a continuous-time invertible generative model with unbiased density estimation and one-pass sampling, while allowing unrestricted neural network architectures. We demonstrate our approach on high-dimensional density estimation, image generation, and variational inference, achieving the state-of-the-art among exact likelihood methods with efficient sampling.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
