DPGN: Distribution Propagation Graph Network for Few-Shot Learning

L. Yang, Liangliang Li, Zilun Zhang, Xinyu Zhou, Erjin Zhou, Yu Liu

2020 · 236 citations · 44 references

Concepts

TL;DR

Graph‑network meta‑learning typically captures only instance‑level relations among examples. This work aims to explicitly model distribution‑level relations between a single example and all others in a 1‑vs‑N setting. The proposed Distribution Propagation Graph Network builds a dual complete graph (point and distribution graphs) and propagates labels through multiple update generations. Experiments on few‑shot benchmarks show DPGN surpasses state‑of‑the‑art by 5–12 % in supervised and 7–13 % in semi‑supervised settings, with code to be released.

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%~12% under supervised setting and 7%~13% under semi-supervised setting. Code will be released.

References

44