Phoneme-Based Transliteration of Foreign Names for OOV Problem

One problem seriously affecting CLIR performance is the processing of queries with embedded foreign names. A proper noun dictionary is never complete rendering name translation from English to Chinese ineffective. One way to solve this problem is not to rely on a dictionary alone but to adopt automatic translation according to pronunciation similarities, i.e. to map phonemes comprising an English name to sound units (e.g. pinyin) of the corresponding Chinese name. This process is called transliteration. We present a statistical transliteration method for CLIR applications. An efficient algorithm for phoneme alignment is described. Unlike traditional rule-based approaches, our method is data-driven. So it is independent of dialect features in Chinese. In addition, it is different from other statistical approaches based on source-channel framework in that we adopt a direct transliteration model, i.e. the direction of probabilistic estimation is consistent with transliteration direction. We demonstrate comparable performance on accuracy to other systems.