Uncertainty decoding for DNN-HMM hybrid systems based on numerical sampling

Hümmer C, Maas R, Schwarz A, Astudillo RF, Kellermann W (2015)


Publication Language: English

Publication Status: Published

Publication Type: Conference contribution, Conference Contribution

Publication year: 2015

Pages Range: 3556-3560

Event location: Dresden DE

ISBN: 978-1-5108-1790-6

Abstract

In this article, we propose an uncertainty decoding scheme for DNN-HMM hybrid systems based on numerical sampling. A finite set of samples is drawn from the estimated probability distribution of the acoustic features and subsequently passed through feature transformations/extensions and the deep neural network (DNN). Then, the nonlinearly-transformed feature samples are averaged at the output of the DNN in order to approximate the posterior distribution of the context-dependent Hidden Markov Model (HMM) states. This concept is experimentally verified for the REVERB challenge task using a reverberation-robust DNN-HMM hybrid system: The numerical sampling is performed in the logmelspec domain, where we estimate the posterior distribution of the acoustic features by combining coherence-based Wiener filtering and uncertainty propagation. The experimental results highlight the good performance of the proposed uncertainty decoding scheme with significantly increased recognition accuracy even for a small number of feature samples.

Authors with CRIS profile

Involved external institutions

How to cite

APA:

Hümmer, C., Maas, R., Schwarz, A., Astudillo, R.F., & Kellermann, W. (2015). Uncertainty decoding for DNN-HMM hybrid systems based on numerical sampling. In Proceedings of the Annual Conference of the International Speech Communication Association (Interspeech) (pp. 3556-3560). Dresden, DE.

MLA:

Hümmer, Christian, et al. "Uncertainty decoding for DNN-HMM hybrid systems based on numerical sampling." Proceedings of the Annual Conference of the International Speech Communication Association (Interspeech), Dresden 2015. 3556-3560.

BibTeX: Download