This study introduces a novel Curriculum Learning based Probabilistic Linear Discriminant Analysis (CL-PLDA) algorithm for improving speaker recognition in noisy conditions. CL-PLDA operates by initializing the training EM algorithm with cleaner data ( easy examples), and successively adds noisier data ( difficult examples) as the training progresses. This curriculum learning based approach guides the parameters of CL-PLDA to better local minima compared to regular PLDA. We test CL-PLDA on speaker verification task of the severely noisy and degraded DARPA RATS data, and show it to significantly outperform regular PLDA across test-sets of varying duration.
Cite as: Ranjan, S., Misra, A., Hansen, J.H.L. (2017) Curriculum Learning Based Probabilistic Linear Discriminant Analysis for Noise Robust Speaker Recognition. Proc. Interspeech 2017, 3717-3721, doi: 10.21437/Interspeech.2017-1199
@inproceedings{ranjan17b_interspeech,
author={Shivesh Ranjan and Abhinav Misra and John H.L. Hansen},
title={{Curriculum Learning Based Probabilistic Linear Discriminant Analysis for Noise Robust Speaker Recognition}},
year=2017,
booktitle={Proc. Interspeech 2017},
pages={3717--3721},
doi={10.21437/Interspeech.2017-1199}
}