Attention Back-end for Automatic Speaker Verification with Multiple Enrollment Utterances

TL;DR

This paper introduces a novel attention-based back-end model for speaker verification that effectively utilizes multiple enrollment utterances, improving accuracy over traditional methods like PLDA and cosine similarity across various datasets.

Contribution

The paper proposes a new attention back-end model employing scaled-dot and feed-forward self-attention networks for better intra-relationship learning among enrollment utterances in speaker verification.

Findings

Lower EER and minDCF scores on CNCeleb with multiple enrollments

Effective for both text-independent and text-dependent verification

Applicable even with a single enrollment utterance

Abstract

Probabilistic linear discriminant analysis (PLDA) or cosine similarity have been widely used in traditional speaker verification systems as back-end techniques to measure pairwise similarities. To make better use of multiple enrollment utterances, we propose a novel attention back-end model, which can be used for both text-independent (TI) and text-dependent (TD) speaker verification, and employ scaled-dot self-attention and feed-forward self-attention networks as architectures that learn the intra-relationships of the enrollment utterances. In order to verify the proposed attention back-end, we conduct a series of experiments on CNCeleb and VoxCeleb datasets by combining it with several sate-of-the-art speaker encoders including TDNN and ResNet. Experimental results using multiple enrollment utterances on CNCeleb show that the proposed attention back-end model leads to lower EER and minDCF score than the PLDA and cosine similarity counterparts for each speaker encoder and an experiment on VoxCeleb indicate that our model can be used even for single enrollment case.