Transfer learning and subword sampling for asymmetric-resource one-to-many neural translation

• Published in: Machine Translation
• Main Authors: Grönroos, Stig-Arne, Virpioja, Sami, Kurimo, Mikko
• Other Authors: Dept Signal Process and Acoust, University of Helsinki, Aalto-yliopisto, Aalto University
• Format: Article in Journal/Newspaper
• Language: English
• Published: Springer Netherlands 2020
• Subjects: Low-resource languages; Multilingual machine translation; Transfer learning; Multi-task learning; Denoising sequence autoencoder; Subword segmentation; North Sámi; Sámi
• Online Access: https://aaltodoc.aalto.fi/handle/123456789/102739; https://doi.org/10.1007/s10590-020-09253-x

| openaire: EC/H2020/780069/EU//MeMAD There are several approaches for improving neural machine translation for low-resource languages: monolingual data can be exploited via pretraining or data augmentation; parallel corpora on related language pairs can be used via parameter sharing or transfer learning in multilingual models; subword segmentation and regularization techniques can be applied to ensure high coverage of the vocabulary. We review these approaches in the context of an asymmetric-resource one-to-many translation task, in which the pair of target languages are related, with one being a very low-resource and the other a higher-resource language. We test various methods on three artificially restricted translation tasks—English to Estonian (low-resource) and Finnish (high-resource), English to Slovak and Czech, English to Danish and Swedish—and one real-world task, Norwegian to North Sámi and Finnish. The experiments show positive effects especially for scheduled multi-task learning, denoising autoencoder, and subword sampling. Peer reviewed