Hybrid Graph Neural Networks for Few-Shot Learning

TL;DR

This paper introduces a hybrid GNN model with two GNNs to improve few-shot learning by better adapting feature embeddings and handling poorly sampled shots, achieving state-of-the-art results.

Contribution

The paper proposes a novel hybrid GNN architecture with two GNNs for improved inductive few-shot learning and robustness to poorly sampled shots.

Findings

Achieves state-of-the-art results on three FSL benchmarks.

Effectively adapts feature embeddings for new tasks.

Handles outliers and distribution overlaps in few-shot samples.

Abstract

Graph neural networks (GNNs) have been used to tackle the few-shot learning (FSL) problem and shown great potentials under the transductive setting. However under the inductive setting, existing GNN based methods are less competitive. This is because they use an instance GNN as a label propagation/classification module, which is jointly meta-learned with a feature embedding network. This design is problematic because the classifier needs to adapt quickly to new tasks while the embedding does not. To overcome this problem, in this paper we propose a novel hybrid GNN (HGNN) model consisting of two GNNs, an instance GNN and a prototype GNN. Instead of label propagation, they act as feature embedding adaptation modules for quick adaptation of the meta-learned feature embedding to new tasks. Importantly they are designed to deal with a fundamental yet often neglected challenge in FSL, that is, with only a handful of shots per class, any few-shot classifier would be sensitive to badly sampled shots which are either outliers or can cause inter-class distribution overlapping. %Our two GNNs are designed to address these two types of poorly sampled few-shots respectively and their complementarity is exploited in the hybrid GNN model. Extensive experiments show that our HGNN obtains new state-of-the-art on three FSL benchmarks.