Discriminatively Re-trained i-vector Extractor for Speaker Recognition [article]

Ondrej Novotny, Oldrich Plchot, Ondrej Glembek, Lukas Burget, Pavel Matejka
2018 arXiv   pre-print
In this work we revisit discriminative training of the i-vector extractor component in the standard speaker verification (SV) system. The motivation of our research lies in the robustness and stability of this large generative model, which we want to preserve, and focus its power towards any intended SV task. We show that after generative initialization of the i-vector extractor, we can further refine it with discriminative training and obtain i-vectors that lead to better performance on
more » ... benchmarks representing different acoustic domains.
arXiv:1810.13183v1 fatcat:gzu5lsm5ybdmnffudpf2ygldai