# The Information-Autoencoding Family: A Lagrangian Perspective on Latent Variable Generative Modeling

Shengjia ZhaoJiaming SongStefano Ermon

A variety of learning objectives have been recently proposed for training generative models. We show that many of them, including InfoGAN, ALI/BiGAN, ALICE, CycleGAN, VAE, $\beta$-VAE, adversarial autoencoders, AVB, and InfoVAE, are Lagrangian duals of the same primal optimization problem... (read more)

PDF Abstract