Partner-Assisted Learning for Few-Shot Image Classification

TL;DR

This paper introduces Partner-Assisted Learning (PAL), a novel two-stage training strategy for few-shot image classification that improves feature extraction and prototype estimation, outperforming existing methods on multiple benchmarks.

Contribution

The paper proposes a new two-stage training scheme, PAL, that enhances feature learning for few-shot classification by modeling pair-wise similarities and aligning features with soft-anchors.

Findings

PAL outperforms state-of-the-art methods on four benchmarks.

Ablation studies validate the effectiveness of each component.

The method improves prototype estimation from limited samples.

Abstract

Few-shot Learning has been studied to mimic human visual capabilities and learn effective models without the need of exhaustive human annotation. Even though the idea of meta-learning for adaptation has dominated the few-shot learning methods, how to train a feature extractor is still a challenge. In this paper, we focus on the design of training strategy to obtain an elemental representation such that the prototype of each novel class can be estimated from a few labeled samples. We propose a two-stage training scheme, Partner-Assisted Learning (PAL), which first trains a partner encoder to model pair-wise similarities and extract features serving as soft-anchors, and then trains a main encoder by aligning its outputs with soft-anchors while attempting to maximize classification performance. Two alignment constraints from logit-level and feature-level are designed individually. For each few-shot task, we perform prototype classification. Our method consistently outperforms the state-of-the-art method on four benchmarks. Detailed ablation studies of PAL are provided to justify the selection of each component involved in training.