On the Efficiency of Integrating Self-supervised Learning and Meta-learning for User-defined Few-shot Keyword Spotting

TL;DR

This paper investigates the combined use of self-supervised learning and meta-learning for user-defined few-shot keyword spotting, finding that HuBERT with Matching network performs best and is robust across different few-shot scenarios.

Contribution

It systematically evaluates the effectiveness of integrating self-supervised learning with meta-learning for few-shot keyword discovery, identifying optimal combinations.

Findings

HuBERT with Matching network achieves top performance.

The combination is robust to variations in few-shot examples.

Self-supervised and meta-learning approaches can be complementary.

Abstract

User-defined keyword spotting is a task to detect new spoken terms defined by users. This can be viewed as a few-shot learning problem since it is unreasonable for users to define their desired keywords by providing many examples. To solve this problem, previous works try to incorporate self-supervised learning models or apply meta-learning algorithms. But it is unclear whether self-supervised learning and meta-learning are complementary and which combination of the two types of approaches is most effective for few-shot keyword discovery. In this work, we systematically study these questions by utilizing various self-supervised learning models and combining them with a wide variety of meta-learning algorithms. Our result shows that HuBERT combined with Matching network achieves the best result and is robust to the changes of few-shot examples.