Adapting Speaker Embeddings for Speaker Diarisation

TL;DR

This paper introduces three techniques to adapt speaker embeddings specifically for speaker diarisation, significantly improving performance by reducing error rates across challenging datasets.

Contribution

It proposes dimensionality reduction, attention-based embedding aggregation, and non-speech clustering to enhance speaker embeddings for diarisation tasks.

Findings

Achieved an average 25.07% reduction in diarisation error rate.

All three proposed techniques contributed positively to performance.

Demonstrated effectiveness on various challenging datasets.

Abstract

The goal of this paper is to adapt speaker embeddings for solving the problem of speaker diarisation. The quality of speaker embeddings is paramount to the performance of speaker diarisation systems. Despite this, prior works in the field have directly used embeddings designed only to be effective on the speaker verification task. In this paper, we propose three techniques that can be used to better adapt the speaker embeddings for diarisation: dimensionality reduction, attention-based embedding aggregation, and non-speech clustering. A wide range of experiments is performed on various challenging datasets. The results demonstrate that all three techniques contribute positively to the performance of the diarisation system achieving an average relative improvement of 25.07% in terms of diarisation error rate over the baseline.