References & Citations
Computer Science > Computation and Language
Title: Character-based NMT with Transformer
(Submitted on 12 Nov 2019)
Abstract: Character-based translation has several appealing advantages, but its performance is in general worse than a carefully tuned BPE baseline. In this paper we study the impact of character-based input and output with the Transformer architecture. In particular, our experiments on EN-DE show that character-based Transformer models are more robust than their BPE counterpart, both when translating noisy text, and when translating text from a different domain. To obtain comparable BLEU scores in clean, in-domain data and close the gap with BPE-based models we use known techniques to train deeper Transformer models.
Submission history
From: Laurent Besacier [view email][v1] Tue, 12 Nov 2019 16:32:38 GMT (2324kb,D)
Link back to: arXiv, form interface, contact.