Title :
Combining Discriminative Feature, Transform, and Model Training for Large Vocabulary Speech Recognition
Author :
Jing Zheng ; Cetin, Omer ; Mei-Yuh Hwang ; Xin Lei ; Stolcke, Andreas ; Morgan, Nigel
Author_Institution :
Lab. of Speech Technol. & Res., SRI Int., Menlo Park, CA, USA
Abstract :
Recent developments in large vocabulary continuous speech recognition (LVCSR) have shown the effectiveness of discriminative training approaches, employing the following three representative techniques: discriminative Gaussian training using the minimum phone error (MPE) criterion, discriminately trained features estimated by multilayer perceptrons (MLPs); and discriminative feature transforms such as feature-level MPE (fMPE). Although MLP features, MPE models, and fMPE transforms have each been shown to improve recognition accuracy, no previous work has applied all three in a single LVCSR system. This paper uses a state-of-the-art Mandarin recognition system as a platform to study the interaction of all three techniques. Experiments in the broadcast news and broadcast conversation domains show that the contribution of each technique is nonredundant, and that the full combination yields the best performance and has good domain generalization.
Keywords :
Gaussian processes; multilayer perceptrons; speech processing; speech recognition; transforms; MLP; Mandarin recognition system; broadcast conversation domains; broadcast news; discriminative Gaussian training; discriminative feature; discriminative training approaches; large vocabulary speech recognition; minimum phone error criterion; model training; multilayer perceptrons; transform; Broadcasting; Computer errors; Computer science; Contacts; Hidden Markov models; Laboratories; Mel frequency cepstral coefficient; Multilayer perceptrons; Speech recognition; Vocabulary; MLP; MPE; Mandarin LVCSR; fMPE;
Conference_Titel :
Acoustics, Speech and Signal Processing, 2007. ICASSP 2007. IEEE International Conference on
Conference_Location :
Honolulu, HI
Print_ISBN :
1-4244-0727-3
DOI :
10.1109/ICASSP.2007.366992