Title :
Using aggregation to improve the performance of mixture Gaussian acoustic models
Author :
Hazen, Timothy J.
Author_Institution :
Lab. for Comput. Sci., MIT, Cambridge, MA
Abstract :
This paper investigates the use of aggregation as a means of improving the performance and robustness of mixture Gaussian models. This technique produces models that are more accurate and more robust to different test sets than traditional cross-validation using a development set. A theoretical justification for this technique is presented along with experimental results in phonetic classification, phonetic recognition, and word recognition tasks on the TIMIT and Resource Management corpora. In speech classification and recognition tasks error rate reductions of up to 12% were observed using this technique. A method for utilizing tree-structured density functions for the purpose of pruning the aggregated models is also presented
Keywords :
Gaussian processes; acoustic signal processing; error statistics; learning (artificial intelligence); pattern classification; probability; speech recognition; Resource Management corpus; TIMIT; aggregated models pruning; aggregation; cross-validation; development set; error rate reduction; experimental results; mixture Gaussian acoustic models; performance; phonetic classification; phonetic recognition; robustness; speech classification; speech recognition; supervised training; test sets; tree-structured density functions; unsupervised training; word recognition; Clustering algorithms; Density functional theory; Error analysis; Laboratories; Management training; Natural languages; Performance evaluation; Resource management; Robustness; Testing;
Conference_Titel :
Acoustics, Speech and Signal Processing, 1998. Proceedings of the 1998 IEEE International Conference on
Conference_Location :
Seattle, WA
Print_ISBN :
0-7803-4428-6
DOI :
10.1109/ICASSP.1998.675349