ISCA Archive Eurospeech 1997
ISCA Archive Eurospeech 1997

A memory management method for a large word network

Tomohiro Iwasaki, Yoshiharu Abe

To improve the performance of continuous speech recognition, it is effective to incorporate grammatical knowledge of task into a word network of a FSN (finite state network) form. But, recently , some of them requires huge memory, so we introduce an ef ficient memory management method for a large word network; distributed FSN model and hiearchical memory model. The system keeps the word network divided to small sub-networks, and activates each sub-network when necessary. Using this method, we can recognize continuously spoken sentences of Japanese addresses, which are made of 390K geographic names, with only 5.6 Mbytes local memory in average.


doi: 10.21437/Eurospeech.1997-70

Cite as: Iwasaki, T., Abe, Y. (1997) A memory management method for a large word network. Proc. 5th European Conference on Speech Communication and Technology (Eurospeech 1997), 171-174, doi: 10.21437/Eurospeech.1997-70

@inproceedings{iwasaki97_eurospeech,
  author={Tomohiro Iwasaki and Yoshiharu Abe},
  title={{A memory management method for a large word network}},
  year=1997,
  booktitle={Proc. 5th European Conference on Speech Communication and Technology (Eurospeech 1997)},
  pages={171--174},
  doi={10.21437/Eurospeech.1997-70},
  issn={1018-4074}
}