Learning to Self-Train for Semi-Supervised Few-Shot Classification

TL;DR

This paper introduces a semi-supervised meta-learning approach called learning to self-train (LST) that improves few-shot classification by intelligently selecting and labeling unlabeled data through iterative self-training and a learned weighting network.

Contribution

The paper proposes a novel semi-supervised meta-learning method that meta-learns to select and weight unlabeled data for improved few-shot classification performance.

Findings

LST significantly outperforms previous methods on ImageNet benchmarks.

The learned weighting network effectively improves pseudo-label quality.

Iterative self-training enhances model performance in semi-supervised few-shot tasks.

Abstract

Few-shot classification (FSC) is challenging due to the scarcity of labeled training data (e.g. only one labeled data point per class). Meta-learning has shown to achieve promising results by learning to initialize a classification model for FSC. In this paper we propose a novel semi-supervised meta-learning method called learning to self-train (LST) that leverages unlabeled data and specifically meta-learns how to cherry-pick and label such unsupervised data to further improve performance. To this end, we train the LST model through a large number of semi-supervised few-shot tasks. On each task, we train a few-shot model to predict pseudo labels for unlabeled data, and then iterate the self-training steps on labeled and pseudo-labeled data with each step followed by fine-tuning. We additionally learn a soft weighting network (SWN) to optimize the self-training weights of pseudo labels so that better ones can contribute more to gradient descent optimization. We evaluate our LST method on two ImageNet benchmarks for semi-supervised few-shot classification and achieve large improvements over the state-of-the-art method. Code is at https://github.com/xinzheli1217/learning-to-self-train.