Noise-Conditioned Mixture-of-Experts Framework for Robust Speaker Verification
IEEE Signal Processing Letters · 2026 · 33 · 1230-1234
DOI: 10.1109/lsp.2026.3673617 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Signal Processing Letters, 2026. Verfasst von Bin Gu und 2 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 28 Literaturangaben. Gefördert von: Provincial Department of Education Program.
Abstract (Originalsprache)
Robust speaker verification under noisy conditions remains an open challenge. Conventional deep learning methods learn a robust unified speaker representation space against diverse background noise and achieve significant improvement. In contrast, this letter presents a noise-conditioned mixture-of-experts framework that decomposes the feature space into specialized noise-aware subspaces for speaker verification. Specifically, we propose a noise-conditioned expert routing mechanism, a universal model based expert specialization strategy, and an SNR-decaying curriculum learning protocol, collectively improving model robustness and generalization under diverse noise conditions. The proposed method can automatically route inputs to expert networks based on noise information derived from the inputs, where each expert targets distinct noise characteristics while preserving speaker identity information. Comprehensive experiments demonstrate consistent superiority over baselines.
Förderung
- Provincial Department of Education Program
Zitieren
APA
Gu, B., Zhao, H., & Wei, J. (2026). Noise-Conditioned Mixture-of-Experts Framework for Robust Speaker Verification. IEEE Signal Processing Letters, 33, 1230-1234. https://doi.org/10.1109/lsp.2026.3673617
Vancouver
Gu B, Zhao H, Wei J. Noise-Conditioned Mixture-of-Experts Framework for Robust Speaker Verification. IEEE Signal Processing Letters. 2026;33:1230-1234. doi:10.1109/lsp.2026.3673617
BibTeX
@article{Gu2026,
author = {Gu, Bin and Zhao, Haitao and Wei, Jibo},
title = {Noise-Conditioned Mixture-of-Experts Framework for Robust Speaker Verification},
journal = {IEEE Signal Processing Letters},
year = {2026},
volume = {33},
pages = {1230--1234},
doi = {10.1109/lsp.2026.3673617}
}
EUVIMED