Partial AUC optimization based deep speaker embeddings with class-center learning for text-independent speaker verification

TL;DR

This paper introduces a novel partial AUC optimization loss function for deep speaker embeddings in text-independent speaker verification, demonstrating competitive performance with existing methods.

Contribution

It proposes a verification-oriented pAUC loss function and a class-center based training method to enhance training efficiency and effectiveness.

Findings

pAUC loss achieves performance comparable to identification loss functions.

Class-center training improves training efficiency for the proposed loss.

Experiments on SITW and NIST SRE 2016 validate the effectiveness of the approach.

Abstract

Deep embedding based text-independent speaker verification has demonstrated superior performance to traditional methods in many challenging scenarios. Its loss functions can be generally categorized into two classes, i.e., verification and identification. The verification loss functions match the pipeline of speaker verification, but their implementations are difficult. Thus, most state-of-the-art deep embedding methods use the identification loss functions with softmax output units or their variants. In this paper, we propose a verification loss function, named the maximization of partial area under the Receiver-operating-characteristic (ROC) curve (pAUC), for deep embedding based text-independent speaker verification. We also propose a class-center based training trial construction method to improve the training efficiency, which is critical for the proposed loss function to be comparable to the identification loss in performance. Experiments on the Speaker in the Wild (SITW) and NIST SRE 2016 datasets show that the proposed pAUC loss function is highly competitive with the state-of-the-art identification loss functions.