Towards Character-Level Transformer NMT by Finetuning Subword Systems

Jindřich Libovický, Alexander Fraser

Applying the Transformer architecture on the character level usually requires very deep architectures that are difficult and slow to train. A few approaches have been proposed that partially overcome this problem by using explicit segmentation into tokens. We show that by initially training a subword model based on this segmentation and then finetuning it on characters, we can obtain a neural machine translation model that works at the character level without requiring segmentation. Without changing the vanilla 6-layer Transformer Base architecture, we train purely character-level models. Our character-level models better capture morphological phenomena and show much higher robustness towards source-side noise at the expense of somewhat worse overall translation quality. Our study is a significant step towards high-performance character-based models that are not extremely large.

Knowledge Graph



Sign up or login to leave a comment