FIGR: Few-shot Image Generation with Reptile

Abstract

Generative Adversarial Networks (GAN) boast impressive capacity to generate realistic images. However, like much of the field of deep learning, they require an inordinate amount of data to produce results, thereby limiting their usefulness in generating novelty. In the same vein, recent advances in meta-learning have opened the door to many few-shot learning applications. In the present work, we propose Few-shot Image Generation using Reptile (FIGR), a GAN meta-trained with Reptile. Our model successfully generates novel images on both MNIST and Omniglot with as little as 4 images from an unseen class. We further contribute FIGR-8, a new dataset for few-shot image generation, which contains 1,548,944 icons categorized in over 18,409 classes. Trained on FIGR-8, initial results show that our model can generalize to more advanced concepts (such as "bird" and "knife") from as few as 8 samples from a previously unseen class of images and as little as 10 training steps through those 8 images. This work demonstrates the potential of training a GAN for few-shot image generation and aims to set a new benchmark for future work in the domain.

9 pages

References in corpus (1)

• Layer Normalization

Cited by in corpus (8)

• DAWSON: A Domain Adaptive Few Shot Generation Framework
• DeltaGAN: Towards Diverse Few-shot Image Generation with Sample-Specific Delta
• Augmentation-Interpolative AutoEncoders for Unsupervised Few-Shot Image Generation
• F2GAN: Fusing-and-Filling GAN for Few-shot Image Generation
• DEff-GAN: Diverse Attribute Transfer for Few-Shot Image Synthesis
• D2C: Diffusion-Denoising Models for Few-shot Conditional Generation
• Meta-Learning Conjugate Priors for Few-Shot Bayesian Optimization
• Impression Space from Deep Template Network