# Prototype Propagation Networks (PPN) for Weakly-supervised Few-shot Learning on Category Graph

## TL;DR

This paper introduces Prototype Propagation Networks (PPN), a novel method that leverages weakly-labeled data and category graphs to improve few-shot learning performance by propagating prototypes across related classes.

## Contribution

The paper proposes PPN, which uses an attention mechanism on category graphs to propagate class prototypes, enhancing few-shot learning with weakly-supervised data.

## Key findings

- PPN outperforms recent few-shot learning methods on benchmarks.
- PPN effectively utilizes weakly-labeled data and category graphs.
- The method is adaptable to different test/inference settings.

## Abstract

A variety of machine learning applications expect to achieve rapid learning from a limited number of labeled data. However, the success of most current models is the result of heavy training on big data. Meta-learning addresses this problem by extracting common knowledge across different tasks that can be quickly adapted to new tasks. However, they do not fully explore weakly-supervised information, which is usually free or cheap to collect. In this paper, we show that weakly-labeled data can significantly improve the performance of meta-learning on few-shot classification. We propose prototype propagation network (PPN) trained on few-shot tasks together with data annotated by coarse-label. Given a category graph of the targeted fine-classes and some weakly-labeled coarse-classes, PPN learns an attention mechanism which propagates the prototype of one class to another on the graph, so that the K-nearest neighbor (KNN) classifier defined on the propagated prototypes results in high accuracy across different few-shot tasks. The training tasks are generated by subgraph sampling, and the training objective is obtained by accumulating the level-wise classification loss on the subgraph. The resulting graph of prototypes can be continually re-used and updated for new tasks and classes. We also introduce two practical test/inference settings which differ according to whether the test task can leverage any weakly-supervised information as in training. On two benchmarks, PPN significantly outperforms most recent few-shot learning methods in different settings, even when they are also allowed to train on weakly-labeled data.