Concept Discovery for Fast Adapatation

TL;DR

This paper introduces COMAML, a concept-based meta-learning approach that enhances few-shot learning by discovering and leveraging structural data features, leading to better adaptation and interpretability.

Contribution

It proposes a novel concept discovery method within meta-learning, improving structural understanding and transferability in few-shot learning tasks.

Findings

COMAML outperforms existing methods on synthetic and real-world datasets.

The approach improves interpretability of meta-learning models.

Enhanced structural feature extraction leads to better adaptation.

Abstract

The advances in deep learning have enabled machine learning methods to outperform human beings in various areas, but it remains a great challenge for a well-trained model to quickly adapt to a new task. One promising solution to realize this goal is through meta-learning, also known as learning to learn, which has achieved promising results in few-shot learning. However, current approaches are still enormously different from human beings' learning process, especially in the ability to extract structural and transferable knowledge. This drawback makes current meta-learning frameworks non-interpretable and hard to extend to more complex tasks. We tackle this problem by introducing concept discovery to the few-shot learning problem, where we achieve more effective adaptation by meta-learning the structure among the data features, leading to a composite representation of the data. Our proposed method Concept-Based Model-Agnostic Meta-Learning (COMAML) has been shown to achieve consistent improvements in the structured data for both synthesized datasets and real-world datasets.