Delving into VoxCeleb: environment invariant speaker recognition

By Joon Son Chung, Jaesung Huh, Seongkyu Mun. 2019

55 citations recorded by the source corpus.

Research in speaker recognition has recently seen significant progress due to the application of neural network models and the availability of new large-scale datasets. There has been a plethora of work in search for more powerful architectures or loss functions suitable for the task, but these works do not consider what information is learnt by the models, apart from being able to predict the given labels. In this work, we introduce an environment adversarial training framework in which the network can effectively learn speaker-discriminative and environment-invariant embeddings without explicit domain shift during training. We achieve this by utilising the previously unused `video' information in the VoxCeleb dataset. The environment adversarial training allows the network to generalise better to unseen conditions. The method is evaluated on both speaker identification and verification tasks using the VoxCeleb dataset, on which we demonstrate significant performance improvements over baselines.

Subfields: Artificial Intelligence, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Jinyu Li
2. Shujie Liu
3. Furu Wei
4. Yu Wu
5. Long Zhou
6. Jaesung Huh
7. Tom Ko
8. Hee-Soo Heo

In Signal Processing

1. Jinyu Li
2. Jaesung Huh
3. Yu Wu
4. Shujie Liu
5. Hee-Soo Heo
6. Bong-Jin Lee
7. Joon Son Chung
8. Tom Ko

Citations in the corpus

• SpeechT5: Unified-Modal Encoder-Decoder Pre-Training for Spoken Language Processing (2021, methodology): “More-over, it outperforms ResNet-based architectures such as Thin ResNet-34 (Chung et al., 2020), indicating the superiority of the encoder-decoder architecture for the SID task.”
• Clova Baseline System for the VoxCeleb Speaker Recognition Challenge 2020 (2020, methodology): “This protocol is in line with that used by [3, 6, 7].”