Generalized End-to-End Loss for Speaker Verification

TL;DR

This paper introduces the generalized end-to-end (GE2E) loss for speaker verification, improving training efficiency and accuracy, and presents the MultiReader technique for domain adaptation across keywords and dialects.

Contribution

The paper proposes the GE2E loss function, which enhances training efficiency and verification accuracy without initial example selection, and introduces MultiReader for domain adaptation.

Findings

GE2E loss reduces speaker verification EER by over 10%

Training time is reduced by 60% with GE2E loss

MultiReader enables accurate multi-keyword and multi-dialect domain adaptation

Abstract

In this paper, we propose a new loss function called generalized end-to-end (GE2E) loss, which makes the training of speaker verification models more efficient than our previous tuple-based end-to-end (TE2E) loss function. Unlike TE2E, the GE2E loss function updates the network in a way that emphasizes examples that are difficult to verify at each step of the training process. Additionally, the GE2E loss does not require an initial stage of example selection. With these properties, our model with the new loss function decreases speaker verification EER by more than 10%, while reducing the training time by 60% at the same time. We also introduce the MultiReader technique, which allows us to do domain adaptation - training a more accurate model that supports multiple keywords (i.e. "OK Google" and "Hey Google") as well as multiple dialects.