SouravDutta91 / map-low-resource-embeddings

Seminar research project: Mapping supervised bilingual word embeddings from English to low-resource languages

Mapping supervised bilingual word embeddings from English to low-resource languages

Seminar research project: Mapping bilingual word embeddings from English to low-resource languages

Here we try to map multilingual embeddings from English to comparatively lower-resource languages in the European region (Estonian, Slovenian, Slovakian, Hungarian).

Disclaimer

This work uses the algorithms mentioned in VecMap (Artetxe et. al. 2018). Credit for the algorithm goes completely to Artexte et. al. I have not changed the algorithms, but used them for low-resource languages to find multilingual embeddings from English to such languages.

Citation

In case you want to cite this work, please use the following BibTex:

@misc{dutta2019mapping,
    title={Mapping Supervised Bilingual Word Embeddings from English to low-resource languages},
    author={Sourav Dutta},
    year={2019},
    eprint={1910.06411},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}