| タイトル | Language Model Combination and Adaptation Using Weighted Finite State Transducers |
| 本文（外部サイト） | http://hdl.handle.net/2060/20110023765 |
| 著者（英） | Liu, X.; Woodland, P. C.; Hieronymus, J. L.; Gales, M. J. F. |
| 著者所属（英） | NASA Ames Research Center |
| 発行日 | 2010-03-15 |
| 言語 | eng |
| 内容記述 | In speech recognition systems language model (LMs) are often constructed by training and combining multiple n-gram models. They can be either used to represent different genres or tasks found in diverse text sources, or capture stochastic properties of different linguistic symbol sequences, for example, syllables and words. Unsupervised LM adaption may also be used to further improve robustness to varying styles or tasks. When using these techniques, extensive software changes are often required. In this paper an alternative and more general approach based on weighted finite state transducers (WFSTs) is investigated for LM combination and adaptation. As it is entirely based on well-defined WFST operations, minimum change to decoding tools is needed. A wide range of LM combination configurations can be flexibly supported. An efficient on-the-fly WFST decoding algorithm is also proposed. Significant error rate gains of 7.3% relative were obtained on a state-of-the-art broadcast audio recognition task using a history dependently adapted multi-level LM modelling both syllable and word sequences |
| NASA分類 | Cybernetics, Artificial Intelligence and Robotics |
| レポートNO | ARC-E-DAA-TN2173 |
| 権利 | Copyright, Distribution under U.S. Government purpose rights |
| URI | https://repository.exst.jaxa.jp/dspace/handle/a-is/244549 |
