[PDF][PDF] ECAPA++: Fine-grained deep embedding learning for TDNN based speaker verification

B Liu, Y Qian - Proc. Annu. Conf. Int. Speech Commun. Assoc, 2023 - isca-archive.org
In this paper, we aim to bridge the performance gap between TDNN and 2D CNN based
speaker verification systems. Specifically, three types of architectural enhancements to …

Towards Lightweight Speaker Verification via Adaptive Neural Network Quantization

B Liu, H Wang, Y Qian - arxiv preprint arxiv:2406.05359, 2024 - arxiv.org
Modern speaker verification (SV) systems typically demand expensive storage and
computing resources, thereby hindering their deployment on mobile devices. In this paper …

Memory-Efficient Training for Deep Speaker Embedding Learning in Speaker Verification

B Liu, Y Qian - arxiv preprint arxiv:2412.01195, 2024 - arxiv.org
Recent speaker verification (SV) systems have shown a trend toward adopting deeper
speaker embedding extractors. Although deeper and larger neural networks can …

Lowbit neural network quantization for speaker verification

H Wang, B Liu, Y Wu, Z Chen… - 2023 IEEE International …, 2023 - ieeexplore.ieee.org
With the continuous development of deep neural networks (DNN) in recent years, the
performance of speaker verification systems has been significantly improved with the …