Contrastive Self-Supervised Learning for Text-Independent Speaker Verification
Authors:
- 6 June 2021
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 6713-6717
Abstract
Current speaker verification models rely on supervised training with massive annotated data. But the collection of labeled utterances from multiple speakers is expensive and facing privacy issues. To open up an opportunity for utilizing massive unlabeled utterance data, our work exploits a contrastive self-supervised learning (CSSL) approach for text-independent speaker verification task. The core principle of CSSL lies in minimizing the distance between the embeddings of augmented segments truncated from the same utterance as well as maximizing those from different utterances. We proposed channel-invariant loss to prevent the network from encoding the undesired channel information into the speaker representation. Bearing these in mind, we conduct intensive experiments on VoxCeleb1&2 datasets. The self-supervised thin-ResNet34 fine-tuned with only 5% of the labeled data can achieve comparable performance to the fully supervised model, which is meaningful to economize lots of manual annotation.Keywords
This publication has 18 references indexed in Scilit:
- SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition2019
- Neural Predictive Coding Using Convolutional Neural Networks Toward Unsupervised Learning of Speaker CharacteristicsIEEE/ACM Transactions on Audio, Speech, and Language Processing, 2019
- Deep Discriminative Embeddings for Duration Robust Speaker Verification2018
- VoxCeleb2: Deep Speaker Recognition2018
- Exploring the Encoding Layer and Loss Function in End-to-End Speaker and Language Recognition System2018
- Boosting Self-Supervised Learning via Knowledge Transfer2018
- X-Vectors: Robust DNN Embeddings for Speaker Recognition2018
- The 2016 NIST Speaker Recognition Evaluation2017
- VoxCeleb: A Large-Scale Speaker Identification Dataset2017
- A study on data augmentation of reverberant speech for robust speech recognition2017