BriefGPT.xyz
Feb, 2020
通过松弛可注入概率流实现规则化自编码器
Regularized Autoencoders via Relaxed Injective Probability Flow
HTML
PDF
Abhishek Kumar, Ben Poole, Kevin Murphy
TL;DR
提出了一种基于概率流的生成模型，通过对模型进行规范化，得到能够提高样本质量的最终目标，进而改进了变分自编码器和自编码器的性能。
Abstract
invertible flow-based generative models
are an effective method for learning to generate samples, while allowing for
tractable likelihood computation
and inference. However, the invertibility requirement restrict
→