Assume, Augment and Learn: Unsupervised Few-Shot Meta-Learning via Random Labels and Data Augmentation

Abstract

The field of few-shot learning has been laboriously explored in the supervised setting, where per-class labels are available. On the other hand, the unsupervised few-shot learning setting, where no labels of any kind are required, has seen little investigation. We propose a method, named Assume, Augment and Learn or AAL, for generating few-shot tasks using unlabeled data. We randomly label a random subset of images from an unlabeled dataset to generate a support set. Then by applying data augmentation on the support set's images, and reusing the support set's labels, we obtain a target set. The resulting few-shot tasks can be used to train any standard meta-learning framework. Once trained, such a model, can be directly applied on small real-labeled datasets without any changes or fine-tuning required. In our experiments, the learned models achieve good generalization performance in a variety of established few-shot learning tasks on Omniglot and Mini-Imagenet.

Work in Progress - Under Review in ICML 2019

References in corpus (2)

• WaveNet: A Generative Model for Raw Audio
• Improved Regularization of Convolutional Neural Networks with Cutout

Cited by in corpus (9)

• Unsupervised Few-shot Learning via Self-supervised Training
• Contrastive Prototype Learning with Augmented Embeddings for Few-Shot Learning
• ProtoTransformer: A Meta-Learning Approach to Providing Student Feedback
• What Can Knowledge Bring to Machine Learning? -- A Survey of Low-shot Learning for Structured Data
• Unsupervised Meta-Learning through Latent-Space Interpolation in Generative Models
• Are Fewer Labels Possible for Few-shot Learning?
• Few Shot Learning With No Labels
• Trainable Class Prototypes for Few-Shot Learning
• Empirical Perspectives on One-Shot Semi-supervised Learning