Happiness is Sharing a Vocabulary: A Study of Transliteration Methods
This work addresses the problem of improving multilingual NLP for languages with non-Latin scripts, though it is incremental as it builds on existing transliteration methods.
The study investigated how shared script, overlapping token vocabularies, and shared phonology affect multilingual model performance in transliteration, finding that romanization significantly outperformed other methods in 7 out of 8 evaluation settings on named entity recognition and natural language inference tasks.
Transliteration has emerged as a promising means to bridge the gap between various languages in multilingual NLP, showing promising results especially for languages using non-Latin scripts. We investigate the degree to which shared script, overlapping token vocabularies, and shared phonology contribute to performance of multilingual models. To this end, we conduct controlled experiments using three kinds of transliteration (romanization, phonemic transcription, and substitution ciphers) as well as orthography. We evaluate each model on two downstream tasks -- named entity recognition (NER) and natural language inference (NLI) -- and find that romanization significantly outperforms other input types in 7 out of 8 evaluation settings, largely consistent with our hypothesis that it is the most effective approach. We further analyze how each factor contributed to the success, and suggest that having longer (subword) tokens shared with pre-trained languages leads to better utilization of the model.