Uniform Sampling over Episode Difficulty
This addresses a core but understudied issue in few-shot learning, offering a simple, algorithm-agnostic improvement for researchers and practitioners.
The paper tackles the problem of how to best sample episodes in episodic training for few-shot learning, finding that uniform sampling over episode difficulty outperforms other schemes like curriculum and easy-/hard-mining, leading to improved accuracies across various datasets, algorithms, and architectures.
Episodic training is a core ingredient of few-shot learning to train models on tasks with limited labelled data. Despite its success, episodic training remains largely understudied, prompting us to ask the question: what is the best way to sample episodes? In this paper, we first propose a method to approximate episode sampling distributions based on their difficulty. Building on this method, we perform an extensive analysis and find that sampling uniformly over episode difficulty outperforms other sampling schemes, including curriculum and easy-/hard-mining. As the proposed sampling method is algorithm agnostic, we can leverage these insights to improve few-shot learning accuracies across many episodic training algorithms. We demonstrate the efficacy of our method across popular few-shot learning datasets, algorithms, network architectures, and protocols.