Compositional Prototypical Networks for Few-Shot Classification

TL;DR

This paper introduces Compositional Prototypical Networks (CPN), a novel approach that learns transferable attribute-based prototypes to improve few-shot classification, achieving state-of-the-art results especially in 5-way 1-shot tasks.

Contribution

It proposes a new method that explicitly learns attribute-based component prototypes and adaptively fuses them with visual prototypes for enhanced few-shot learning.

Findings

Achieves state-of-the-art results on multiple datasets.

Significant performance improvements in 5-way 1-shot classification.

Component prototypes demonstrate high transferability.

Abstract

It is assumed that pre-training provides the feature extractor with strong class transferability and that high novel class generalization can be achieved by simply reusing the transferable feature extractor. In this work, our motivation is to explicitly learn some fine-grained and transferable meta-knowledge so that feature reusability can be further improved. Concretely, inspired by the fact that humans can use learned concepts or components to help them recognize novel classes, we propose Compositional Prototypical Networks (CPN) to learn a transferable prototype for each human-annotated attribute, which we call a component prototype. We empirically demonstrate that the learned component prototypes have good class transferability and can be reused to construct compositional prototypes for novel classes. Then a learnable weight generator is utilized to adaptively fuse the compositional and visual prototypes. Extensive experiments demonstrate that our method can achieve state-of-the-art results on different datasets and settings. The performance gains are especially remarkable in the 5-way 1-shot setting. The code is available at https://github.com/fikry102/CPN.