English

Back-ends Selection for Deep Speaker Embeddings

Sound 2022-04-26 v1 Audio and Speech Processing

Abstract

Probabilistic Linear Discriminant Analysis (PLDA) was the dominant and necessary back-end for early speaker recognition approaches, like i-vector and x-vector. However, with the development of neural networks and margin-based loss functions, we can obtain deep speaker embeddings (DSEs), which have advantages of increased inter-class separation and smaller intra-class distances. In this case, PLDA seems unnecessary or even counterproductive for the discriminative embeddings, and cosine similarity scoring (Cos) achieves better performance than PLDA in some situations. Motivated by this, in this paper, we systematically explore how to select back-ends (Cos or PLDA) for deep speaker embeddings to achieve better performance in different situations. By analyzing PLDA and the properties of DSEs extracted from models with different numbers of segment-level layers, we make the conjecture that Cos is better in same-domain situations and PLDA is better in cross-domain situations. We conduct experiments on VoxCeleb and NIST SRE datasets in four application situations, single-/multi-domain training and same-/cross-domain test, to validate our conjecture and briefly explain why back-ends adaption algorithms work.

Keywords

Cite

@article{arxiv.2204.11403,
  title  = {Back-ends Selection for Deep Speaker Embeddings},
  author = {Zhuo Li and Runqiu Xiao and Zihan Zhang and Zhenduo Zhao and Wenchao Wang and Pengyuan Zhang},
  journal= {arXiv preprint arXiv:2204.11403},
  year   = {2022}
}

Comments

submitted to interspeech2022

R2 v1 2026-06-24T10:57:18.410Z