A New Method to Reduce the Ambiguity of Japanese Phoneme Candidates Recognized by Recurrent Neural Networks

Shin-ichiro Hashimukai, Chikahiro Araki, Mikio Mori, Shuji Taniguchi, Shozo Kato, Yasuhiro Ogoshi. A New Method to Reduce the Ambiguity of Japanese Phoneme Candidates Recognized by Recurrent Neural Networks. In ISUC 2008, Second International Symposium on Universal Communication, Osaka, Japan, 15-16 December 2008. pages 353-357, IEEE Computer Society, 2008. [doi]

Abstract

Abstract is missing.