PixelGAN Autoencoders

1. Introduction This paper proposed a “PixelGAN Autoencoder”, for which the generative path is a convolutional autoregressive neural network on pixels, conditioned on a latent code, and the recognition path uses a generative adversarial network (GAN) to impose a prior distribution on the latent code. This paper also shows different priors result in different decompositions […]

Continue reading


Kullback-Leibler Divergence Explained

Introduction This blog is an introduction on the KL-divergence, aka relative entropy. The blog gives a simple example for understand relative entropy, and therefore I will not attempt to re-write the authors words. What I will do, in addition to reading the blog (which can be found at https://www.countbayesie.com/blog/2017/5/9/kullback-leibler-divergence-explained), is try to convey some extra […]

Continue reading


A Sneak Peak of Our Upcoming “AI Tech Report”

Synced has recently begun putting together a research report on the technologies powered Artificial Intelligence (AI) to identify their development paths. You may sign up to receive the report as soon as the rolling out starts. The Background In the report, the concept of Artificial Intelligence (AI) will be represented together with a brief history […]

Continue reading


Deep Learning Paper Sparks Online Feud!

Feature image is created by Jannoon028 – Freepik.com Researchers Yoav Goldberg and Yann Lecun face off on Natural Language Processing Social media is humanity’s new intellectual battlefield. Sports fans, social justice warriors, and even the President of the United States tweet, take to discussion boards or make memes to mock, preach, thrust and parry against […]

Continue reading


Adversarial Generator-Encoder Networks

1. Introduction This paper proposes a new autoencoder-like generative network, called Adversarial Generator-Encoder Networks (AGE Network). This model is special because this AGE does not have any discriminators, which makes the entire architecture much simpler than some recently-proposed GANs, but with nearly the same-level performance of sample generation. 2. Methods As the figure above shows, […]

Continue reading


Unsupervised Image-to-Image Translation with Generative Adversarial Networks

Paper Source: https://arxiv.org/pdf/1701.02676.pdf 1. Introduction This paper proposes a general approach to achieve “Image-to-Image Translation” by using deep convolutional and conditional generative adversarial networks (GANs). In this work, the authors develop a two-step unsupervised learning method to translate images without specifying any correspondence between them. What is “Image-to-Image Translation”? “Image-to-Image Translation” means to transform an […]

Continue reading


ArtGAN – Artwork Synthesis with Conditional Categorical GANs

Paper Source: https://pdfs.semanticscholar.org/ec94/874d38378f53319d467412a124809542d3db.pdf?_ga=1.46132652.922857708.1488461012 Paper Authors: 1. Introduction This paper presents a so-called ArtGAN to generate complex images like artworks as shown in the following figure. Prior works based on GANs are usually used to generate images, which normally have clear, distinguishable foregrounds and backgrounds while often only having one or two (main) objects in each […]

Continue reading