Auxiliary Deep Generative Models, Lars Maaløe, Casper Kaae Sønderby, Søren Kaae Sønderby, Ole Winther, 2016Proceedings of The 33rd International Conference on Machine Learning, Vol. 48 (PMLR) - 介绍了在推断网络中使用辅助变量的概念,以提高VAE中近似后验分布的灵活性和表达能力。
Semi-Amortized Variational Autoencoders, Dongho Kim, Wonkyung Kim, Jeongwoo Kim, Daeho Kim and Seung-won Hwang, 2018Proceedings of the 2nd Workshop on Deep Learning Approaches for Unsupervised and Semi-Supervised Learning (DLASS 2018), Vol. 86 (Proceedings of Machine Learning Research)DOI: 10.48550/arXiv.1805.02100 - 提出并形式化了半摊销变分自编码器框架,将摊销推断与针对每个实例的优化相结合,以改进后验近似。