Statistical Models for Unsupervised, Semi-Supervised, and Supervised Transliteration Mining

<p dir="ltr">We present a generative model that efficiently mines transliteration pairs in a consistent fashion in three different settings: unsupervised, semi-supervised, and supervised transliteration mining. The model interpolates two sub-models, one for the generation of transliteration pairs and one for the generation of non-transliteration pairs (i.e., noise). The model is trained on noisy unlabeled data using the EM algorithm. During training the transliteration sub-model learns to generate transliteration pairs and the fixed non-transliteration model generates the noise pairs. After training, the unlabeled data is disambiguated based on the posterior probabilities of the two sub-models. We evaluate our transliteration mining system on data from a transliteration mining shared task and on parallel corpora. For three out of four language pairs, our system outperforms all semi-supervised and supervised systems that participated in the NEWS 2010 shared task. On word pairs extracted from parallel corpora with fewer than 2% transliteration pairs, our system achieves up to 86.7% F-measure with 77.9% precision and 97.8% recall.</p><h2>Other Information</h2><p dir="ltr">Published in: Computational Linguistics<br>License: <a href="https://creativecommons.org/licenses/by-nc-nd/4.0/" rel="noreferrer noopener" target="_blank">https://creativecommons.org/licenses/by-nc-nd/4.0/</a> <br>See article on publisher's website: <a href="https://dx.doi.org/10.1162/coli_a_00286" target="_blank">https://dx.doi.org/10.1162/coli_a_00286</a></p>