ISCA Archive Interspeech 2008
ISCA Archive Interspeech 2008

Towards vocabulary-independent speech indexing for large-scale repositories

Jian Shao, Roger Peng Yu, Qingwei Zhao, Yonghong Yan, Frank Seide

The Out-Of-Vocabulary problem remains a challenge for wordlattice- based speech indexing. Sub-word-based approaches address this problem effectively for small-scale tasks, but suffer from poor precisions on large-scale databases due to lack of strong language model constraints. We propose a method for searching OOV queries with large-scale databases in two steps. First, result candidates are extracted from a sub-word-based system, ensuring a high recall. The candidates are then refined by word-lattice rescoring aiming at a high precision. Experiments on a 160-hours lecture set show that the proposed approach achieves a relative improvement of 8.7% over the sub-word-based baseline, and 19.7% for only single-word queries.

doi: 10.21437/Interspeech.2008-563

Cite as: Shao, J., Yu, R.P., Zhao, Q., Yan, Y., Seide, F. (2008) Towards vocabulary-independent speech indexing for large-scale repositories. Proc. Interspeech 2008, 2150-2153, doi: 10.21437/Interspeech.2008-563

  author={Jian Shao and Roger Peng Yu and Qingwei Zhao and Yonghong Yan and Frank Seide},
  title={{Towards vocabulary-independent speech indexing for large-scale repositories}},
  booktitle={Proc. Interspeech 2008},