Abstract
Orthographic similarities across languages provide a strong signal for unsupervised probabilistic transduction (decipherment) for closely related language pairs. The existing decipherment models, however, are not well suited for exploiting these orthographic similarities. We propose a log-linear model with latent variables that incorporates orthographic similarity features. Maximum likelihood training is computationally expensive for the proposed log-linear model. To address this challenge, we perform approximate inference via Markov chain Monte Carlo sampling and contrastive divergence. Our results show that the proposed log-linear model with contrastive divergence outperforms the existing generative decipherment models by exploiting the orthographic features. The model both scales to large vocabularies and preserves accuracy in low- and no-resource contexts.- Anthology ID:
- J18-3006
- Volume:
- Computational Linguistics, Volume 44, Issue 3 - September 2018
- Month:
- September
- Year:
- 2018
- Address:
- Cambridge, MA
- Venue:
- CL
- SIG:
- Publisher:
- MIT Press
- Note:
- Pages:
- 525–546
- Language:
- URL:
- https://aclanthology.org/J18-3006
- DOI:
- 10.1162/coli_a_00326
- Cite (ACL):
- Iftekhar Naim, Parker Riley, and Daniel Gildea. 2018. Feature-Based Decipherment for Machine Translation. Computational Linguistics, 44(3):525–546.
- Cite (Informal):
- Feature-Based Decipherment for Machine Translation (Naim et al., CL 2018)
- PDF:
- https://preview.aclanthology.org/ingestion-script-update/J18-3006.pdf