Curriculum learning based approach for noise robust language identification using DNN with attention
作者：
Highlights：
• Curriculum Learning approach for noise robust language identification system.
• DNN and DNN With Attention are used for developing LID systems.
• Experimental verification is carried out using IIIT-H and AP17-OLR databases.
• Various curriculum learning strategies are explored to train multi-SNR models.
• Proposed models have performed better in terms of equal error rate.
摘要
•Curriculum Learning approach for noise robust language identification system.•DNN and DNN With Attention are used for developing LID systems.•Experimental verification is carried out using IIIT-H and AP17-OLR databases.•Various curriculum learning strategies are explored to train multi-SNR models.•Proposed models have performed better in terms of equal error rate.
论文关键词：Automatic language identification,Background environments,I-vector,Deep neural network (DNN),DNN with attention (DNN-WA),Multi signal-to-noise (SNR) models,Curriculum learning
论文评审过程：Received 6 December 2017, Revised 11 May 2018, Accepted 2 June 2018, Available online 7 June 2018, Version of Record 18 June 2018.
论文官网地址：https://doi.org/10.1016/j.eswa.2018.06.004
