On the Copying Problem of Unsupervised NMT: A Training Schedule with a Language Discriminator Loss
This addresses a specific issue in machine translation for low-resource languages, but it is incremental as it builds on existing unsupervised methods.
The paper tackled the copying problem in unsupervised neural machine translation, where parts of the input are copied as output, especially for distant or low-resource language pairs, and proposed a training schedule with a language discriminator loss that improved translation performance on low-resource languages.
Although unsupervised neural machine translation (UNMT) has achieved success in many language pairs, the copying problem, i.e., directly copying some parts of the input sentence as the translation, is common among distant language pairs, especially when low-resource languages are involved. We find this issue is closely related to an unexpected copying behavior during online back-translation (BT). In this work, we propose a simple but effective training schedule that incorporates a language discriminator loss. The loss imposes constraints on the intermediate translation so that the translation is in the desired language. By conducting extensive experiments on different language pairs, including similar and distant, high and low-resource languages, we find that our method alleviates the copying problem, thus improving the translation performance on low-resource languages.