DAWSON: A Domain Adaptive Few Shot Generation Framework

TL;DR

DAWSON is a flexible meta-learning framework that enables GANs to adapt quickly to new domains with minimal data, demonstrated through applications in music and digit generation.

Contribution

It introduces a novel training procedure for meta-learning GANs, supporting various architectures and algorithms, and presents the first few-shot music generation model.

Findings

Music generation adapts with only tens of songs.

MNIST digit generation with only four samples.

Framework supports broad meta-learning and GAN variants.

Abstract

Training a Generative Adversarial Networks (GAN) for a new domain from scratch requires an enormous amount of training data and days of training time. To this end, we propose DAWSON, a Domain Adaptive FewShot Generation FrameworkFor GANs based on meta-learning. A major challenge of applying meta-learning GANs is to obtain gradients for the generator from evaluating it on development sets due to the likelihood-free nature of GANs. To address this challenge, we propose an alternative GAN training procedure that naturally combines the two-step training procedure of GANs and the two-step training procedure of meta-learning algorithms. DAWSON is a plug-and-play framework that supports a broad family of meta-learning algorithms and various GANs with architectural-variants. Based on DAWSON, We also propose MUSIC MATINEE, which is the first few-shot music generation model. Our experiments show that MUSIC MATINEE could quickly adapt to new domains with only tens of songs from the target domains. We also show that DAWSON can learn to generate new digits with only four samples in the MNIST dataset. We release source codes implementation of DAWSON in both PyTorch and Tensorflow, generated music samples on two genres and the lightning video.