Incorrect citation for CVAE

[june6423]

There are two different papers out there called CVAEs. They work in completely different ways, but people get confused and often cite them incorrectly.

First paper :Kingma et al. "Semi-supervised Learning with Deep Generative Models", NIPS 2014

Second paper : Sohn et al, "Learning Structured Output Representation using Deep Conditional Generative Models", NIPS 2015

The first paper came out first, and given the look of the method, it's natural to call it CVAE. However, the first paper refers to itself as a conditional generative model (M2), and the term CVAE never appears.

The first use of the term CVAE is in the second paper. The second paper clearly states that his methos is a CVAE.

Despite the first use of the term CVAE in the second paper, the first paper is often cited as CVAE. Since Kingma, the author of the first paper, is the author of VAE, it is more natural to refer to the first paper as CVAE.

CVAE has been used interchangeably to refer to both papers. However, this github implementation is closer to the first paper's implementation.