Few-Shot Learning with Global Class Representations

TL;DR

This paper introduces a novel few-shot learning method that learns global class representations using both base and novel samples, employing a registration module and sample synthesis to improve generalization.

Contribution

It proposes a joint training approach with a registration module and sample synthesis, enabling effective learning of global class representations for both standard and generalized FSL.

Findings

Effective in standard FSL setting

Extends well to generalized FSL

Outperforms existing methods

Abstract

In this paper, we propose to tackle the challenging few-shot learning (FSL) problem by learning global class representations using both base and novel class training samples. In each training episode, an episodic class mean computed from a support set is registered with the global representation via a registration module. This produces a registered global class representation for computing the classification loss using a query set. Though following a similar episodic training pipeline as existing meta learning based approaches, our method differs significantly in that novel class training samples are involved in the training from the beginning. To compensate for the lack of novel class training samples, an effective sample synthesis strategy is developed to avoid overfitting. Importantly, by joint base-novel class training, our approach can be easily extended to a more practical yet challenging FSL setting, i.e., generalized FSL, where the label space of test data is extended to both base and novel classes. Extensive experiments show that our approach is effective for both of the two FSL settings.