Fully Unsupervised Training of Few-shot Keyword Spotting

TL;DR

This paper introduces an unsupervised, synthetic-data-based approach for few-shot keyword spotting that leverages metric learning and speech synthesis to eliminate the need for labeled datasets.

Contribution

It presents a fully unsupervised FS-KWS system trained solely on synthetic speech data using metric learning and speech synthesis with pseudo phonemes.

Findings

Competitive performance on real datasets without labeled data

Effective use of synthetic multi-view samples for training

Elimination of the need for large labeled datasets

Abstract

For training a few-shot keyword spotting (FS-KWS) model, a large labeled dataset containing massive target keywords has known to be essential to generalize to arbitrary target keywords with only a few enrollment samples. To alleviate the expensive data collection with labeling, in this paper, we propose a novel FS-KWS system trained only on synthetic data. The proposed system is based on metric learning enabling target keywords to be detected using distance metrics. Exploiting the speech synthesis model that generates speech with pseudo phonemes instead of texts, we easily obtain a large collection of multi-view samples with the same semantics. These samples are sufficient for training, considering metric learning does not intrinsically necessitate labeled data. All of the components in our framework do not require any supervision, making our method unsupervised. Experimental results on real datasets show our proposed method is competitive even without any labeled and real datasets.