Abstract
Neural machine translation decoders are usually conditional language models to sequentially generate words for target sentences. This approach is limited to find the best word composition and requires help of explicit methods as beam search. To help learning correct compositional mechanisms in NMTs, we propose concept equalization using direct mapping distributed representations of source and target sentences. In a translation experiment from English to French, the concept equalization significantly improved translation quality by 3.00 BLEU points compared to a state-of-the-art NMT model.- Anthology ID:
- I17-2051
- Volume:
- Proceedings of the Eighth International Joint Conference on Natural Language Processing (Volume 2: Short Papers)
- Month:
- November
- Year:
- 2017
- Address:
- Taipei, Taiwan
- Venue:
- IJCNLP
- SIG:
- Publisher:
- Asian Federation of Natural Language Processing
- Note:
- Pages:
- 302–307
- Language:
- URL:
- https://aclanthology.org/I17-2051
- DOI:
- Cite (ACL):
- Kangil Kim, Jong-Hun Shin, Seung-Hoon Na, and SangKeun Jung. 2017. Concept Equalization to Guide Correct Training of Neural Machine Translation. In Proceedings of the Eighth International Joint Conference on Natural Language Processing (Volume 2: Short Papers), pages 302–307, Taipei, Taiwan. Asian Federation of Natural Language Processing.
- Cite (Informal):
- Concept Equalization to Guide Correct Training of Neural Machine Translation (Kim et al., IJCNLP 2017)
- PDF:
- https://preview.aclanthology.org/nodalida-main-page/I17-2051.pdf