Abstract / Description of output
Incorporating source syntactic information into neural machine translation (NMT) has recently proven successful (Eriguchi et al., 2016; Luong et al., 2016). However, this is generally done using an outside parser to syntactically annotate the training data, making this technique difficult to use for languages or domains for which a reliable parser is not available. In this paper, we introduce an unsupervised tree-to-sequence (tree2seq) model for neural machine translation; this model is able to induce an unsupervised hierarchical structure on the source sentence based on the downstream task of neural machine translation. We adapt the Gumbel tree-LSTM of Choi et al. (2018) to NMT in order to create the encoder.
We evaluate our model against sequential and supervised parsing baselines on three low- and medium-resource language pairs. For low-resource cases, the unsupervised tree2seq encoder significantly outperforms the baselines; no improvements are seen for medium-resource translation.
We evaluate our model against sequential and supervised parsing baselines on three low- and medium-resource language pairs. For low-resource cases, the unsupervised tree2seq encoder significantly outperforms the baselines; no improvements are seen for medium-resource translation.
| Original language | English |
|---|---|
| Title of host publication | Proceedings of the Workshop on the Relevance of Linguistic Structure in Neural Architectures for NLP |
| Place of Publication | Melbourne, Australia |
| Publisher | ACL Anthology |
| Pages | 6-12 |
| Number of pages | 7 |
| Publication status | Published - Jul 2018 |
| Event | NLP Workshop 2018 - Melbourne, Australia Duration: 19 Jul 2018 → 19 Jul 2018 https://sites.google.com/view/relsnnlp/home |
Conference
| Conference | NLP Workshop 2018 |
|---|---|
| Country/Territory | Australia |
| City | Melbourne |
| Period | 19/07/18 → 19/07/18 |
| Internet address |