Directed graphical models of classifier combination: application to phone recognition.
作者： Katrin Kirchhoff , Jeff A. Bilmes
DOI:
关键词:
摘要: Classifier combination is a technique that often provides appreciable accuracy gains. In this paper, we argue the underlying statistical model of classifier should be made explicit. Using directed graphical models (DGMs), provide representations two common schemes, mean and product rules. We also introduce new DGMs yield novel find these DGM-inspired rules can achieve significant gains on TIMIT phone-classification task relative to existing schemes.
isca-speech.org参考文章(27)
Saeed Vaseghi, Paul M. McCourt, Philip McMahon, Discriminative weighting of multi-resolution sub-band cepstral features for speech recognition conference of the international speech communication association. ,(1998)
Amanda J. C. Sharkey, Multi-Net Systems Perspectives in Neural Computing. pp. 1- 30 ,(1999) , 10.1007/978-1-4471-0793-4_1
Nikki Mirghafori, Nelson Morgan, Combining connectionist multi-band and full-band probability streams for speech recognition of natural numbers. conference of the international speech communication association. ,(1998)
J. Bilmes, N. Morgan, Su-Lin Wu, H. Bourlard, Stochastic perceptual speech models with durational dependence international conference on spoken language processing. ,vol. 3, pp. 1301- 1304 ,(1996) , 10.1109/ICSLP.1996.607851
Ross D. Shachter, Bayes-ball: Rational pastime (for determining irrelevance and requisite information in belief networks and influence diagrams) uncertainty in artificial intelligence. pp. 480- 487 ,(1998)
Christopher M. Bishop, Neural networks for pattern recognition ,(1995)
Katrin Kirchhoff, Combining articulatory and acoustic information for speech recognition in noisy and reverberant environments conference of the international speech communication association. ,(1998)
Andrew K. Halberstadt, James R. Glass, Heterogeneous measurements and multiple classifiers for speech recognition. conference of the international speech communication association. ,(1998)
Nir Friedman, Dan Geiger, Moises Goldszmidt, Bayesian Network Classifiers Machine Learning. ,vol. 29, pp. 131- 163 ,(1997) , 10.1023/A:1007465528199
J.G. Fiscus, A post-processing system to yield reduced word error rates: Recognizer Output Voting Error Reduction (ROVER) ieee automatic speech recognition and understanding workshop. pp. 347- 354 ,(1997) , 10.1109/ASRU.1997.659110