Massively Multilingual Neural Grapheme-to-Phoneme Conversion
This addresses the challenge of building text-to-speech and speech recognition systems for low-resource languages, offering a more scalable and compact solution compared to traditional monolingual approaches.
The authors tackled the problem of grapheme-to-phoneme conversion for low-resource languages by developing a massively multilingual neural sequence-to-sequence model trained on hundreds of languages, achieving an 11% improvement in phoneme error rate over monolingual adaptation methods.
Grapheme-to-phoneme conversion (g2p) is necessary for text-to-speech and automatic speech recognition systems. Most g2p systems are monolingual: they require language-specific data or handcrafting of rules. Such systems are difficult to extend to low resource languages, for which data and handcrafted rules are not available. As an alternative, we present a neural sequence-to-sequence approach to g2p which is trained on spelling--pronunciation pairs in hundreds of languages. The system shares a single encoder and decoder across all languages, allowing it to utilize the intrinsic similarities between different writing systems. We show an 11% improvement in phoneme error rate over an approach based on adapting high-resource monolingual g2p models to low-resource languages. Our model is also much more compact relative to previous approaches.