Learning to Pronounce Chinese Without a Pronunciation Dictionary

TL;DR

This paper presents an unsupervised approach for converting Chinese characters to Mandarin speech, achieving high accuracy without relying on a pronunciation dictionary, surpassing previous methods.

Contribution

It introduces a novel unsupervised method to learn character-to-syllable mappings from non-parallel data, enabling pronunciation without a dictionary.

Findings

Character-to-syllable accuracy of 89%

Outperforms prior work with 22% accuracy

Effective decoding of Chinese writing into speech

Abstract

We demonstrate a program that learns to pronounce Chinese text in Mandarin, without a pronunciation dictionary. From non-parallel streams of Chinese characters and Chinese pinyin syllables, it establishes a many-to-many mapping between characters and pronunciations. Using unsupervised methods, the program effectively deciphers writing into speech. Its token-level character-to-syllable accuracy is 89%, which significantly exceeds the 22% accuracy of prior work.