Noise adaptive training for robust automatic speech recognition

Authors:
Ozlem Kalinli;Michael L. Seltzer;Jasha Droppo;Alex Acero
Affiliations:
The R & D Group, Sony Computer Entertainment of America, Foster City, CA;Microsoft Research, Redmond, WA;Microsoft Research, Redmond, WA;Microsoft Research, Redmond, WA
Venue:
IEEE Transactions on Audio, Speech, and Language Processing
Year:
2010

Citing 3
Cited 3

Speech recognition in noisy environments using first-order vector Taylor series

Speech Communication
Acoustical and Environmental Robustness in Automatic Speech Recognition

Acoustical and Environmental Robustness in Automatic Speech Recognition
A vector Taylor series approach for environment-independent speech recognition

ICASSP '96 Proceedings of the Acoustics, Speech, and Signal Processing, 1996. on Conference Proceedings., 1996 IEEE International Conference - Volume 02

Recognition of consonant-vowel (CV) units under background noise using combined temporal and spectral preprocessing

International Journal of Speech Technology
Noise-robust speech recognition through auditory feature detection and spike sequence decoding

Neural Computation
A MAP-based Online Estimation Approach to Ensemble Speaker and Speaking Environment Modeling

IEEE/ACM Transactions on Audio, Speech and Language Processing (TASLP)

Quantified Score

Hi-index	0.00

Visualization

Abstract

In traditional methods for noise robust automatic speech recognition, the acoustic models are typically trained using clean speech or using multi-condition data that is processed by the same feature enhancement algorithm expected to be used in decoding. In this paper, we propose a noise adaptive training (NAT) algorithm that can be applied to all training data that normalizes the environmental distortion as part of the model training. In contrast to feature enhancement methods, NAT estimates the underlying "pseudo-clean" model parameters directly without relying on point estimates of the clean speech features as an intermediate step. The pseudo-clean model parameters learned with NAT are later used with vector Taylor series (VTS) model adaptation for decoding noisy utterances at test time. Experiments performed on the Aurora 2 and Aurora 3 tasks demonstrate that the proposed NAT method obtain relative improvements of 18.83% and 32.02%, respectively, over VTS model adaptation.