Automatic Allocation of Training Data for Speech Understanding Based on Multiple Model Combinations

Kazunori KOMATANI  Mikio NAKANO  Masaki KATSUMARU  Kotaro FUNAKOSHI  Tetsuya OGATA  Hiroshi G. OKUNO  

Publication
IEICE TRANSACTIONS on Information and Systems   Vol.E95-D   No.9   pp.2298-2307
Publication Date: 2012/09/01
Online ISSN: 1745-1361
DOI: 10.1587/transinf.E95.D.2298
Print ISSN: 0916-8532
Type of Manuscript: PAPER
Category: Speech and Hearing
Keyword: 
spoken dialogue system,  language understanding,  rapid prototyping,  limited amount of training data,  

Full Text: PDF(549.2KB)
>>Buy this Article


Summary: 
The optimal way to build speech understanding modules depends on the amount of training data available. When only a small amount of training data is available, effective allocation of the data is crucial to preventing overfitting of statistical methods. We have developed a method for allocating a limited amount of training data in accordance with the amount available. Our method exploits rule-based methods for when the amount of data is small, which are included in our speech understanding framework based on multiple model combinations, i.e., multiple automatic speech recognition (ASR) modules and multiple language understanding (LU) modules, and then allocates training data preferentially to the modules that dominate the overall performance of speech understanding. Experimental evaluation showed that our allocation method consistently outperforms baseline methods that use a single ASR module and a single LU module while the amount of training data increases.