DPGN: Distribution Propagation Graph Network for Few-shot Learning

TL;DR

DPGN introduces a dual graph network that models both distribution-level and instance-level relations to improve few-shot learning, significantly outperforming existing methods on benchmark datasets.

Contribution

The paper proposes a novel dual graph architecture that explicitly models distribution-level relations alongside instance-level relations for few-shot learning.

Findings

Outperforms state-of-the-art methods by 5-12% in supervised settings.

Achieves 7-13% improvement in semi-supervised few-shot learning.

Demonstrates effectiveness across multiple benchmark datasets.

Abstract

Most graph-network-based meta-learning approaches model instance-level relation of examples. We extend this idea further to explicitly model the distribution-level relation of one example to all other examples in a 1-vs-N manner. We propose a novel approach named distribution propagation graph network (DPGN) for few-shot learning. It conveys both the distribution-level relations and instance-level relations in each few-shot learning task. To combine the distribution-level relations and instance-level relations for all examples, we construct a dual complete graph network which consists of a point graph and a distribution graph with each node standing for an example. Equipped with dual graph architecture, DPGN propagates label information from labeled examples to unlabeled examples within several update generations. In extensive experiments on few-shot learning benchmarks, DPGN outperforms state-of-the-art results by a large margin in 5% $\sim$ 12% under supervised setting and 7% $\sim$ 13% under semi-supervised setting. Code will be released.