Probabilistic and Bottle-Neck Features for LVCSR of Meetings

In recent years, probabilistic features became an integral part of state-of-the-are LVCSR systems. In this work, we are exploring the possibility of obtaining the features directly from neural net without the necessity of converting output probabilities to features suitable for subsequent GMM-HMM sy...

Full description

Saved in:
Bibliographic Details
Main Authors: Grezl, F., Karafiat, M., Kontar, S., Cernocky, J.
Format: Conference Proceeding
Language:English
Subjects:
bottle-neck features
Discrete cosine transforms
Discrete transforms
Feature extraction
Hidden Markov models
LVCSR
meeting recognition
Merging
Neural networks
NIST
Performance evaluation
Principal component analysis
Probabilistic features
Speech recognition
TRAP-based features
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In recent years, probabilistic features became an integral part of state-of-the-are LVCSR systems. In this work, we are exploring the possibility of obtaining the features directly from neural net without the necessity of converting output probabilities to features suitable for subsequent GMM-HMM system. We experimented with 5-layer MLP with bottle-neck in the middle layer. After training such a neural net, we used outputs of the bottle-neck as features for GMM-HMM recognition system. The benefits are twofold: first, improvement was gained when these features are used instead of the probabilistic features, second, the size of the system was reduced, as only part of the neural net is used. The experiments were performed on meetings recognition task defined in MST RT'05 evaluation.
ISSN:1520-6149
2379-190X
DOI:10.1109/ICASSP.2007.367023