On Episodes, Prototypical Networks, and Few-Shot LearningDownload PDF

Published: 09 Nov 2021, Last Modified: 17 Sept 2023NeurIPS 2021 PosterReaders: Everyone
Keywords: meta-learning, few-shot learning, episodic-learning, analysis, case-study
TL;DR: We show that using episodes during training with Matching and Prototypical Networks is a data-inefficient way of exploiting training batches, and can be replaced by simpler non-episodic counterparts to improve performance.
Abstract: Episodic learning is a popular practice among researchers and practitioners interested in few-shot learning. It consists of organising training in a series of learning problems (or episodes), each divided into a small training and validation subset to mimic the circumstances encountered during evaluation. But is this always necessary? In this paper, we investigate the usefulness of episodic learning in methods which use nonparametric approaches, such as nearest neighbours, at the level of the episode. For these methods, we not only show how the constraints imposed by episodic learning are not necessary, but that they in fact lead to a data-inefficient way of exploiting training batches. We conduct a wide range of ablative experiments with Matching and Prototypical Networks, two of the most popular methods that use nonparametric approaches at the level of the episode. Their "non-episodic'' counterparts are considerably simpler, have less hyperparameters, and improve their performance in multiple few-shot classification datasets.
Supplementary Material: pdf
Code Of Conduct: I certify that all co-authors of this work have read and commit to adhering to the NeurIPS Statement on Ethics, Fairness, Inclusivity, and Code of Conduct.
Community Implementations: [![CatalyzeX](/images/catalyzex_icon.svg) 1 code implementation](https://www.catalyzex.com/paper/arxiv:2012.09831/code)
Code: https://github.com/fiveai/on-episodes-fsl
14 Replies