Flipped-Adversarial AutoEncoders

TL;DR

The paper introduces Flipped-Adversarial AutoEncoders, a novel model that improves data reconstruction quality and semantic encoding by combining adversarial training in data space with re-encoding in latent space.

Contribution

It presents a new autoencoder framework that minimizes re-encoding errors and uses adversarial training in data space, differing from prior hybrid approaches.

Findings

Produces sharper reconstructed images

Enables rich semantic data representations

Balances adversarial training with re-encoding errors

Abstract

We propose a flipped-Adversarial AutoEncoder (FAAE) that simultaneously trains a generative model G that maps an arbitrary latent code distribution to a data distribution and an encoder E that embodies an "inverse mapping" that encodes a data sample into a latent code vector. Unlike previous hybrid approaches that leverage adversarial training criterion in constructing autoencoders, FAAE minimizes re-encoding errors in the latent space and exploits adversarial criterion in the data space. Experimental evaluations demonstrate that the proposed framework produces sharper reconstructed images while at the same time enabling inference that captures rich semantic representation of data.