File Download
Supplementary
-
Citations:
- Appears in Collections:
Conference Paper: Meta-learning for low-resource neural machine translation
Title | Meta-learning for low-resource neural machine translation |
---|---|
Authors | |
Issue Date | 2018 |
Publisher | Association for Computational Linguistics. |
Citation | Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP), Brussels, Belgium, 31 October - 4 November 2018, p. 3622-3631 How to Cite? |
Abstract | In this paper, we propose to extend the recently introduced model-agnostic meta-learning algorithm (MAML, Finn, et al., 2017) for low-resource neural machine translation (NMT). We frame low-resource translation as a meta-learning problem where we learn to adapt to low-resource languages based on multilingual high-resource language tasks. We use the universal lexical representation (Gu et al., 2018b) to overcome the input-output mismatch across different languages. We evaluate the proposed meta-learning strategy using eighteen European languages (Bg, Cs, Da, De, El, Es, Et, Fr, Hu, It, Lt, Nl, Pl, Pt, Sk, Sl, Sv and Ru) as source tasks and five diverse languages (Ro,Lv, Fi, Tr and Ko) as target tasks. We show that the proposed approach significantly outperforms the multilingual, transfer learning based approach (Zoph et al., 2016) and enables us to train a competitive NMT system with only a fraction of training examples. For instance, the proposed approach can achieve as high as 22.04 BLEU on Romanian-English WMT’16 by seeing only 16,000 translated words (~600 parallel sentences) |
Persistent Identifier | http://hdl.handle.net/10722/278334 |
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Gu, J | - |
dc.contributor.author | Wang, Y | - |
dc.contributor.author | Chen, Y | - |
dc.contributor.author | Li, VOK | - |
dc.contributor.author | Cho, K | - |
dc.date.accessioned | 2019-10-04T08:11:59Z | - |
dc.date.available | 2019-10-04T08:11:59Z | - |
dc.date.issued | 2018 | - |
dc.identifier.citation | Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP), Brussels, Belgium, 31 October - 4 November 2018, p. 3622-3631 | - |
dc.identifier.uri | http://hdl.handle.net/10722/278334 | - |
dc.description.abstract | In this paper, we propose to extend the recently introduced model-agnostic meta-learning algorithm (MAML, Finn, et al., 2017) for low-resource neural machine translation (NMT). We frame low-resource translation as a meta-learning problem where we learn to adapt to low-resource languages based on multilingual high-resource language tasks. We use the universal lexical representation (Gu et al., 2018b) to overcome the input-output mismatch across different languages. We evaluate the proposed meta-learning strategy using eighteen European languages (Bg, Cs, Da, De, El, Es, Et, Fr, Hu, It, Lt, Nl, Pl, Pt, Sk, Sl, Sv and Ru) as source tasks and five diverse languages (Ro,Lv, Fi, Tr and Ko) as target tasks. We show that the proposed approach significantly outperforms the multilingual, transfer learning based approach (Zoph et al., 2016) and enables us to train a competitive NMT system with only a fraction of training examples. For instance, the proposed approach can achieve as high as 22.04 BLEU on Romanian-English WMT’16 by seeing only 16,000 translated words (~600 parallel sentences) | - |
dc.language | eng | - |
dc.publisher | Association for Computational Linguistics. | - |
dc.relation.ispartof | Conference on Empirical Methods in Natural Language Processing (EMNLP) Proceedings | - |
dc.title | Meta-learning for low-resource neural machine translation | - |
dc.type | Conference_Paper | - |
dc.identifier.email | Li, VOK: vli@eee.hku.hk | - |
dc.identifier.authority | Li, VOK=rp00150 | - |
dc.description.nature | link_to_OA_fulltext | - |
dc.identifier.doi | 10.18653/v1/D18-1398 | - |
dc.identifier.hkuros | 306537 | - |
dc.identifier.spage | 3622 | - |
dc.identifier.epage | 3631 | - |