文献类型：期刊文献

英文题名：Fast Adaptive Meta-Learning for Few-Shot Image Generation

作者：Phaphuangwittayakul, Aniwat[1];Guo, Yi[1,2,3];Ying, Fangli[1,4]

机构：[1]East China Univ Sci & Technol, Dept Comp Sci & Engn, Shanghai 200237, Peoples R China;[2]Natl Engn Lab Big Data Distribut & Exchange Techn, Business Intelligence & Visualisat Res Ctr, Shanghai 200436, Peoples R China;[3]Shanghai Engn Res Ctr Big Data & Internet Audienc, Shanghai 200272, Peoples R China;[4]East China Univ Sci & Technol, State Key Lab Bioreactor Engn, Shanghai 200237, Peoples R China

年份：2022

卷号：24

起止页码：2205

外文期刊名：IEEE TRANSACTIONS ON MULTIMEDIA

收录：;EI(收录号：20212010372379);WOS:【SCI-EXPANDED(收录号:WOS:000778959200033)】；

基金：This work was supported in part by the National Key Research and Development Program of China under Grant 2018YFC0807105, in part by the Science and Technology Committee of Shanghai Municipality (STCSM) under Grants 17DZ1101003, 18511106602, and 18DZ2252300, in part by the Open Funding Project of the State Key Laboratory of Bioreactor Engineering, East China University of Science and Technology, Shanghai, China, and in part by the International College of Digital Innovation (ICDI), Chiang Mai University, Thailand. (Aniwat Phaphuangwittayakul and Fangli Ying contributed equally to this work.)

语种：英文

外文关键词：Task analysis; Training; Image synthesis; Adaptation models; Generative adversarial networks; Generators; Data models; Meta-learning; few-shot image generation; generative adversarial network; unsupervised learning

摘要：Generative Adversarial Networks (GANs) are capable of effectively synthesising new realistic images and estimating the potential distribution of samples utilising adversarial learning. Nevertheless, conventional GANs require a large amount of training data samples to produce plausible results. Inspired by the capacity for humans to quickly learn new concepts from a small number of examples, several meta-learning approaches for the few-shot datasets are presented. However, most of meta-learning algorithms are designed to tackle few-shot classification and reinforcement learning tasks. Moreover, the existing meta-learning models for image generation are complex, thereby affecting the length of training time required. Fast Adaptive Meta-Learning (FAML) based on GAN and the encoder network is proposed in this study for few-shot image generation. This model demonstrates the capability to generate new realistic images from previously unseen target classes with only a small number of examples required. With 10 times faster convergence, FAML requires only one-fourth of the trainable parameters in comparison baseline models by training a simpler network with conditional feature vectors from the encoder, while increasing the number of generator iterations. The visualisation results are demonstrated in the paper. This model is able to improve few-shot image generation with the lowest FID score, highest IS, and comparable LPIPS to MNIST, Omniglot, VGG-Faces, and miniImageNet datasets. The source code is available on https://github.com/phaphuang/FAML.