Efficient Conditional GAN Transfer with Knowledge Propagation across Classes

TL;DR

This paper proposes a novel method for transferring knowledge in conditional GANs by propagating class-specific information through conditional batch normalization, enabling efficient learning of new classes with minimal additional parameters.

Contribution

It introduces a new GAN transfer technique that explicitly propagates knowledge across classes using conditional batch normalization, improving efficiency and performance.

Findings

Outperforms state-of-the-art methods in conditional GAN transfer tasks

Enables linear growth of BN parameters with new classes

Demonstrates effective knowledge sharing among classes

Abstract

Generative adversarial networks (GANs) have shown impressive results in both unconditional and conditional image generation. In recent literature, it is shown that pre-trained GANs, on a different dataset, can be transferred to improve the image generation from a small target data. The same, however, has not been well-studied in the case of conditional GANs (cGANs), which provides new opportunities for knowledge transfer compared to unconditional setup. In particular, the new classes may borrow knowledge from the related old classes, or share knowledge among themselves to improve the training. This motivates us to study the problem of efficient conditional GAN transfer with knowledge propagation across classes. To address this problem, we introduce a new GAN transfer method to explicitly propagate the knowledge from the old classes to the new classes. The key idea is to enforce the popularly used conditional batch normalization (BN) to learn the class-specific information of the new classes from that of the old classes, with implicit knowledge sharing among the new ones. This allows for an efficient knowledge propagation from the old classes to the new ones, with the BN parameters increasing linearly with the number of new classes. The extensive evaluation demonstrates the clear superiority of the proposed method over state-of-the-art competitors for efficient conditional GAN transfer tasks. The code is available at: https://github.com/mshahbazi72/cGANTransfer