Evolutionary Generative Adversarial Networks with Crossover Based Knowledge Distillation

TL;DR

This paper introduces CE-GAN, an evolutionary GAN framework that employs crossover and knowledge distillation to improve training stability and image quality, addressing common GAN issues like mode collapse.

Contribution

It proposes a novel crossover operator and integrates knowledge distillation into evolutionary GANs, enhancing performance and training efficiency.

Findings

CE-GAN produces higher quality generated images.

The method improves training stability and reduces mode collapse.

CE-GAN is more time-efficient than traditional GANs.

Abstract

Generative Adversarial Networks (GAN) is an adversarial model, and it has been demonstrated to be effective for various generative tasks. However, GAN and its variants also suffer from many training problems, such as mode collapse and gradient vanish. In this paper, we firstly propose a general crossover operator, which can be widely applied to GANs using evolutionary strategies. Then we design an evolutionary GAN framework C-GAN based on it. And we combine the crossover operator with evolutionary generative adversarial networks (EGAN) to implement the evolutionary generative adversarial networks with crossover (CE-GAN). Under the premise that a variety of loss functions are used as mutation operators to generate mutation individuals, we evaluate the generated samples and allow the mutation individuals to learn experiences from the output in a knowledge distillation manner, imitating the best output outcome, resulting in better offspring. Then, we greedily selected the best offspring as parents for subsequent training using discriminator as evaluator. Experiments on real datasets demonstrate the effectiveness of CE-GAN and show that our method is competitive in terms of generated images quality and time efficiency.