Teaching keyword spotters to spot new keywords with limited examples

TL;DR

This paper introduces KeySEM, a speech embedding model pre-trained on keyword recognition, enabling rapid and effective learning of new keywords from limited examples, suitable for personalized and on-device keyword spotting.

Contribution

We propose KeySEM, a novel pre-trained speech embedding model that improves few-shot keyword learning and generalizes across languages without re-training on previous keywords.

Findings

KeySEM outperforms existing methods with fewer examples.

It generalizes well to multiple languages.

It allows sequential learning of new keywords without re-training.

Abstract

Learning to recognize new keywords with just a few examples is essential for personalizing keyword spotting (KWS) models to a user's choice of keywords. However, modern KWS models are typically trained on large datasets and restricted to a small vocabulary of keywords, limiting their transferability to a broad range of unseen keywords. Towards easily customizable KWS models, we present KeySEM (Keyword Speech EMbedding), a speech embedding model pre-trained on the task of recognizing a large number of keywords. Speech representations offered by KeySEM are highly effective for learning new keywords from a limited number of examples. Comparisons with a diverse range of related work across several datasets show that our method achieves consistently superior performance with fewer training examples. Although KeySEM was pre-trained only on English utterances, the performance gains also extend to datasets from four other languages indicating that KeySEM learns useful representations well aligned with the task of keyword spotting. Finally, we demonstrate KeySEM's ability to learn new keywords sequentially without requiring to re-train on previously learned keywords. Our experimental observations suggest that KeySEM is well suited to on-device environments where post-deployment learning and ease of customization are often desirable.