Title :
Combination of Acoustic Classifiers Based on Dempster-Shafer Theory of Evidence
Author :
Valente, Fabio ; Hermansky, Hynek
Author_Institution :
IDIAP Res. Inst.
Abstract :
In this paper we investigate combination of neural net based classifiers using Dempster-Shafer theory of evidence. Under some assumptions, combination rule resembles a product of errors rule observed in human speech perception. Different combination are tested in ASR experiments both in matched and mismatched conditions and compared with more conventional probability combination rules. Proposed techniques are particularly effective in mismatched conditions.
Keywords :
speech processing; speech recognition; acoustic classifiers; automatic speech recognition; evidence Dempster-Shafer theory; human speech perception; neural net based classifiers; Atomic measurements; Automatic speech recognition; Bayesian methods; Humans; Multilayer perceptrons; Neural networks; Speech recognition; Testing; Classifier combination; Dempster-Shafer theory; Multi-Stream ASR;
Conference_Titel :
Acoustics, Speech and Signal Processing, 2007. ICASSP 2007. IEEE International Conference on
Conference_Location :
Honolulu, HI
Print_ISBN :
1-4244-0727-3
Electronic_ISBN :
1520-6149
DOI :
10.1109/ICASSP.2007.367273
