GLeaD: Improving GANs with A Generator-Leading Task

TL;DR

This paper introduces a novel GAN training paradigm where the discriminator is tasked with aiding the generator by extracting features that G can decode, leading to improved image generation quality.

Contribution

The paper proposes a generator-leading task for the discriminator in GANs, fostering better collaboration and significantly enhancing generation quality compared to traditional methods.

Findings

FID improved from 4.30 to 2.55 on LSUN Bedroom

FID improved from 4.04 to 2.82 on LSUN Church

Demonstrates substantial performance gains over baselines

Abstract

Generative adversarial network (GAN) is formulated as a two-player game between a generator (G) and a discriminator (D), where D is asked to differentiate whether an image comes from real data or is produced by G. Under such a formulation, D plays as the rule maker and hence tends to dominate the competition. Towards a fairer game in GANs, we propose a new paradigm for adversarial training, which makes G assign a task to D as well. Specifically, given an image, we expect D to extract representative features that can be adequately decoded by G to reconstruct the input. That way, instead of learning freely, D is urged to align with the view of G for domain classification. Experimental results on various datasets demonstrate the substantial superiority of our approach over the baselines. For instance, we improve the FID of StyleGAN2 from 4.30 to 2.55 on LSUN Bedroom and from 4.04 to 2.82 on LSUN Church. We believe that the pioneering attempt present in this work could inspire the community with better designed generator-leading tasks for GAN improvement.