Revisiting hybrid and GMM-HMM system combination techniques
- 1 May 2013
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 15206149,p. 6744-6748
Abstract
In this paper we investigate techniques to combine hybrid HMM-DNN (hidden Markov model - deep neural network) and tandem HMM-GMM (hidden Markov model - Gaussian mixture model) acoustic models using: (1) model averaging, and (2) lattice combination with Minimum Bayes Risk decoding. We have performed experiments on the “TED Talks” task following the protocol of the IWSLT-2012 evaluation. Our experimental results suggest that DNN-based and GMM-based acoustic models are complementary, with error rates being reduced by up to 8% relative when the DNN and GMM systems are combined at model-level in a multi-pass automatic speech recognition (ASR) system. Additionally, further gains were obtained by combining model-averaged lattices with the one obtained from baseline systems.Keywords
This publication has 29 references indexed in Scilit:
- Minimum Bayes Risk decoding and system combination based on a recursion for edit distanceComputer Speech & Language, 2011
- Directed decision trees for generating complementary systemsSpeech Communication, 2009
- A Fast Learning Algorithm for Deep Belief NetsNeural Computation, 2006
- Product of Gaussians for speech recognitionComputer Speech & Language, 2006
- Finding consensus in speech recognition: word error minimization and other applications of confusion networksComputer Speech & Language, 2000
- Maximum likelihood linear transformations for HMM-based speech recognitionComputer Speech & Language, 1998
- Multilayer perceptrons for state-dependent weightings of HMM likelihoodsSpeech Communication, 1996
- Connectionist probability estimators in HMM speech recognitionIEEE Transactions on Speech and Audio Processing, 1994
- A hybrid segmental neural net/hidden Markov model system for continuous speech recognitionIEEE Transactions on Speech and Audio Processing, 1994
- Combining TDNN and HMM in a hybrid system for improved continuous-speech recognitionIEEE Transactions on Speech and Audio Processing, 1994