K. Kirchhoff and J.A. Bilmes. Combination and joint training of acoustic classifiers for speech recognition. ISCA ITRW Workshop on Automatic Speech Recognition - Challenges for the new millenium (ASRU2000), pages 17-23, 2000. Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
Concluding Remarks on Multi-band and Multi-stream Research for.. - Hagen (2002) (Correct)
.... have often shown that, despite the limitations of the inaccurate independence assumption between the different recognizers working on each combination of subbands, the recombination by a product can be a more effective method of combining the outputs of multiple classifiers than the sum rule [6, 1, 5, 7, 10, 11]. FC product rules for likelihoods Under the inaccurate assumption of independence between the different recognizers, the full likelihood can be decomposed into a product of B stream likelihoods for each state qk, according to: P(xlqk) Ok H pW(xilqk) 7) with p(xilqk) the state likelihood of ....
K. Kirchhoff and J.A. Bilmes. Combination and joint training of acoustic classifiers for speech recognition. ISCA ITRW Workshop on Automatic Speech Recognition - Challenges for the new millenium (ASRU2000), pages 17-23, 2000.
A Comparison And Combination Of Methods For OOV Word Detection .. - Hazen, Bazzi (2001) (4 citations) (Correct)
....accept reject classifier such as a neural network [9, 12] 4. COMBINING OOV WORD DETECTION AND CONFIDENCE SCORING In speech recognition research, it has been discovered that combining the outputs of different classifiers and or recognizers can improve recognition accuracy and robustness [2, 4, 8]. These results are most compelling when the different classifiers utilize different observation measurements or modeling approaches but achieve similar results. Under these circumstances, the expected gain from combining the different classifiers is the greatest. This was the motivation for ....
K. Kirchhoff and J. Bilmes, "Combination and joint training of acoustic classifiers for speech recognition," Proc. of ISCA ASR2000.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC