Applying the Transformer to Character-level Transduction
This work addresses a specific bottleneck in NLP for researchers and practitioners by enabling better performance on character-level tasks, though it is incremental as it builds on existing transformer methods.
The paper tackled the problem of applying the transformer model to character-level transduction tasks, where it previously underperformed compared to recurrent models, and found that using a large batch size enables the transformer to achieve state-of-the-art performance on tasks like morphological inflection and historical text normalization, with improvements shown on grapheme-to-phoneme conversion and transliteration.
The transformer has been shown to outperform recurrent neural network-based sequence-to-sequence models in various word-level NLP tasks. Yet for character-level transduction tasks, e.g. morphological inflection generation and historical text normalization, there are few works that outperform recurrent models using the transformer. In an empirical study, we uncover that, in contrast to recurrent sequence-to-sequence models, the batch size plays a crucial role in the performance of the transformer on character-level tasks, and we show that with a large enough batch size, the transformer does indeed outperform recurrent models. We also introduce a simple technique to handle feature-guided character-level transduction that further improves performance. With these insights, we achieve state-of-the-art performance on morphological inflection and historical text normalization. We also show that the transformer outperforms a strong baseline on two other character-level transduction tasks: grapheme-to-phoneme conversion and transliteration.