Abstract
We study several methods for full or partial sharing of the decoder parameters of multi-lingual NMT models. Using only the WMT 2019 shared task parallel datasets for training, we evaluate both fully supervised and zero-shot translation performance in 110 unique translation directions. We use additional test sets and re-purpose evaluation methods recently used for unsupervised MT in order to evaluate zero-shot translation performance for language pairs where no gold-standard parallel data is available. To our knowledge, this is the largest evaluation of multi-lingual translation yet conducted in terms of the total size of the training data we use, and in terms of the number of zero-shot translation pairs we evaluate. We conduct an in-depth evaluation of the translation performance of different models, highlighting the trade-offs between methods of sharing decoder parameters. We find that models which have task-specific decoder parameters outperform models where decoder parameters are fully shared across all tasks.- Anthology ID:
- W19-5319
- Volume:
- Proceedings of the Fourth Conference on Machine Translation (Volume 2: Shared Task Papers, Day 1)
- Month:
- August
- Year:
- 2019
- Address:
- Florence, Italy
- Venue:
- WMT
- SIG:
- SIGMT
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 209–217
- Language:
- URL:
- https://aclanthology.org/W19-5319
- DOI:
- 10.18653/v1/W19-5319
- Cite (ACL):
- Chris Hokamp, John Glover, and Demian Gholipour Ghalandari. 2019. Evaluating the Supervised and Zero-shot Performance of Multi-lingual Translation Models. In Proceedings of the Fourth Conference on Machine Translation (Volume 2: Shared Task Papers, Day 1), pages 209–217, Florence, Italy. Association for Computational Linguistics.
- Cite (Informal):
- Evaluating the Supervised and Zero-shot Performance of Multi-lingual Translation Models (Hokamp et al., WMT 2019)
- PDF:
- https://preview.aclanthology.org/starsem-semeval-split/W19-5319.pdf