Generalized End-to-End Loss for Speaker Verification
Authors:
Top Cited Papers
- 1 April 2018
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 4879-4883
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.Keywords
All Related Versions
This publication has 9 references indexed in Scilit:
- End-to-end text-dependent speaker verification2016
- FaceNet: A unified embedding for face recognition and clustering2015
- Automatic gain control and multi-style training for robust small-footprint keyword spotting with deep neural networks2015
- Deep neural networks for small footprint text-dependent speaker verification2014
- Small-footprint keyword spotting using deep neural networks2014
- Front-End Factor Analysis for Speaker VerificationIEEE Transactions on Audio, Speech, and Language Processing, 2010
- An overview of text-independent speaker recognition: From features to supervectorsSpeech Communication, 2010
- A Tutorial on Text-Independent Speaker VerificationEURASIP Journal on Advances in Signal Processing, 2004
- Long Short-Term MemoryNeural Computation, 1997