日本語の未知語に対する発音付与のための多対多アライメント

IR IPSJ HANDLE Web Site Open Access

Bibliographic Information

Other Title
  • ニホンゴ ノ ミチゴ ニ タイスル ハツオン フヨ ノ タメ ノ タタイ タアライメント
  • Many-to-many Alignment Algorithm for Automatic Pronunciation Annotation on Japanese Unknown Words

Search this article

Abstract

The need for robust pronunciation annotation on unknown words has been increasing with the development of an application that deals with proper nouns and brand-new words, such as Spoken Document Retrieval and Voice Search. In robust pronunciation annotation on unknown words, the alignment between graphemes and phonemes is vital data. In this paper, for the purpose of the improving pronunciation annotation on Japanese unknown words, we propose the alignment algorithm that requires a mapping with small unit having high expression ability on unknown words while avoiding degradation of the accuracy of a mapping between graphemes and phonemes. An evaluation experiment of a many-to-many alignment by automatic pronunciation annotation using Web text mining is also performed. That experimental result shows that the proposed many-to-many alignment obtains 3.9 point improvement on recall rate that represents the generalization ability for unknown words while avoiding degradation of the accuracy of the pronunciation annotation compared with the conventional approach.

Journal

Related Projects

See more

Keywords

Details 詳細情報について

Report a problem

Back to top