Издательство — Речевые технологии — Выпуск №2/2008

The paper deals with automatic diacritization in systems of automatic speech recognition for the Arabic languages. The authors give an outline of the major difficulties in Arabic the researchers usually face when they start working on the problem of automatic diacritization. This part is fol lowed by re review of methods and approaches that have been developed. The final part is devoted to the newly developed buckWalter algorithm which allows removing incorrectly diacritized words from the dictionaries.