THESIS
2018
ix, 41 pages : illustrations ; 30 cm
Abstract
Few-shot learning aims to enable machine learning models to learn new concepts from
few labeled instances. In this thesis, we propose a conceptually simple and general framework
called MetaGAN for few-shot learning problems. Most state-of-the-art few-shot
classification models can be integrated with MetaGAN in a principled and straightforward
way. By introducing an adversarial generator conditioned on tasks, we augment
vanilla few-shot classification models with the ability to discriminate between real and
fake data. We argue that this GAN-based approach can help few-shot classifiers to learn
sharper decision boundary, which could generalize better. We show that with our MetaGAN framework, we can extend supervised few-shot learning models to naturally cope
with unlabeled data. D...[
Read more ]
Few-shot learning aims to enable machine learning models to learn new concepts from
few labeled instances. In this thesis, we propose a conceptually simple and general framework
called MetaGAN for few-shot learning problems. Most state-of-the-art few-shot
classification models can be integrated with MetaGAN in a principled and straightforward
way. By introducing an adversarial generator conditioned on tasks, we augment
vanilla few-shot classification models with the ability to discriminate between real and
fake data. We argue that this GAN-based approach can help few-shot classifiers to learn
sharper decision boundary, which could generalize better. We show that with our MetaGAN framework, we can extend supervised few-shot learning models to naturally cope
with unlabeled data. Different from previous work in semi-supervised few-shot learning,
our algorithms can deal with semi-supervision at both sample-level and task-level. We
give theoretical justifications of the strength of MetaGAN, and validate the effectiveness
of MetaGAN on challenging few-shot image classification benchmarks.
Post a Comment