MineGAN++: Mining Generative Models for Efficient Knowledge Transfer to Limited Data Domains

TL;DR

MineGAN++ introduces a novel knowledge transfer approach for GANs that mines beneficial knowledge from pretrained models, improving generation quality in limited data domains while avoiding common pitfalls like mode collapse.

Contribution

The paper presents MineGAN++, a new method that mines and transfers the most relevant knowledge from multiple pretrained GANs to target domains with limited data, enhancing transfer efficiency.

Findings

Outperforms existing transfer methods on challenging datasets

Effectively transfers from multiple pretrained GANs

Reduces overfitting with sparse subnetwork selection

Abstract

GANs largely increases the potential impact of generative models. Therefore, we propose a novel knowledge transfer method for generative models based on mining the knowledge that is most beneficial to a specific target domain, either from a single or multiple pretrained GANs. This is done using a miner network that identifies which part of the generative distribution of each pretrained GAN outputs samples closest to the target domain. Mining effectively steers GAN sampling towards suitable regions of the latent space, which facilitates the posterior finetuning and avoids pathologies of other methods, such as mode collapse and lack of flexibility. Furthermore, to prevent overfitting on small target domains, we introduce sparse subnetwork selection, that restricts the set of trainable neurons to those that are relevant for the target dataset. We perform comprehensive experiments on several challenging datasets using various GAN architectures (BigGAN, Progressive GAN, and StyleGAN) and show that the proposed method, called MineGAN, effectively transfers knowledge to domains with few target images, outperforming existing methods. In addition, MineGAN can successfully transfer knowledge from multiple pretrained GANs.