Mova0.5 pre-release⌕
☰
home / docs / papers / Semitic languages / Morphology-rich Alphasyllabary Embeddings

Morphology-rich Alphasyllabary Embeddings

What it is about

The authors train word embeddings for Amharic, a Semitic language written in the Fidel abugida, and compare them with Arabic. They collect about 16 million tokens of Amharic news and a new word-analogy set made by native speakers (1,731 analogies, of which 1,163 are morphological and 568 semantic). Three pre-processing options are tried with fastText-style subword models: keep the Fidel characters, alphabetize (split each character into consonant and vowel), and abjadize (drop the vowels). A consonant-vowel neural model is also tried. The best morphological accuracy is for the alphabetized subword model (49.85), ahead of the Fidel model (37.64), the abjadized model (26.87) and plain skip-gram (8.44). For semantic analogies, plain skip-gram is best (5.87), and every subword model is lower. The authors also sweep the co-occurrence order, which trades morphological against semantic accuracy, and show that the Semitic embeddings are less affected by it than the others.

Ideas

For Mova

I1 is a design rule for the unpointed mode of tools/hbo and a note in Phonology of Biblical Hebrew. I2 is a test set for tools/hbo and 05. The verb: system, binyanim, tense and use, built from MACULA. I3 is a reporting rule for any vector component of world.

Cite

The paper itself, not this card. BibTeX from the library's record of the work.

@inproceedings{mersha2020morphology,
  title = {{Morphology-rich Alphasyllabary Embeddings}},
  author = {Mersha, Amanuel and Wu, Stephen},
  booktitle = {Proceedings of the Twelfth Language Resources and Evaluation Conference},
  year = {2020},
  url = {https://aclanthology.org/2020.lrec-1.315/}
}

References

19 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Abdou, M., Kulmizev, A., and Ravishankar, V. (2018). Mgad: Multilingual generation of analogy datasets. In Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC-2018). find
  2. Ammar, W., Mulcaire, G., Tsvetkov, Y., Lample, G., Dyer, C., and Smith, N. A. (2016). Massively multilingual word embeddings. arXiv preprint arXiv:1602.01925. ↗
  3. Artetxe, M., Labaka, G., Lopez-Gazpio, I., and Agirre, E. (2018). Uncovering divergent linguistic information in word embeddings with lessons for intrinsic and extrinsic evaluation. Proceedings of the 22nd Conference on the Computational Natural Language Learning (CoNLL), pages 282–291. find
  4. Blei, D. M., Ng, A. Y., and Jordan, M. I. (2003). Latent dirichlet allocation. Journal of Machine Learning Research, 3:993–1022. find
  5. Bojanowski, P., Grave, E., Joulin, A., and Mikolov, T. (2017). Enriching word vectors with subword information. Transactions of the Association for Computational Linguistics, 5:135–146. find
  6. Chen, X., Xu, L., Liu, Z., Sun, M., and Luan, H.-B. (2015). Joint learning of character and word embeddings. In IJCAI, pages 1236–1242. find
  7. Dahou, A., Xiong, S., Zhou, J., Haddoud, M. H., and Duan, P. (2016). Word embeddings and convolutional neural network for arabic sentiment classification. In Proceedings of COLING 2016, the 26th International Conference on Computational Linguistics: Technical Papers, pages 2418–2427. find
  8. Deerwester, S., Dumais, S., Furnas, G. W., Landauer, T. K., and Harshman, R. (1990). Indexing by latent semantic analysis. Journal of the American Society for Information Science, 41(6):391–407. find
  9. Eberhard, D. M., Simons, G. F., and Fennig, C. D. (2020). Ethnologue: Languages of the World. SIL International, Dallas, Texas, 23 edition. find
  10. Grave, E., Bojanowski, P., Gupta, P., Joulin, A., and Mikolov, T. (2018). Learning word vectors for 157 languages. In Proceedings of the International Conference on Language Resources and Evaluation (LREC 2018). find
  11. Jaech, A., Mulcaire, G., Hathi, S., Ostendorf, M., and Smith, N. A. (2016). Hierarchical character-word models for language identification. In Conference on Empirical Methods in Natural Language Processing, page 84. find
  12. Kim, Y., Jernite, Y., Sontag, D., and Rush, A. M. (2016). Character-aware neural language models. In AAAI, pages 2741–2749. find
  13. Levy, O. and Goldberg, Y. (2014). Linguistic regularities in sparse and explicit word representations. In Proceedings of the eighteenth conference on computational natural language learning, pages 171–180. find
  14. Mikolov, T., Chen, K., Corrado, G., and Dean, J. (2013). Efficient estimation of word representations in vector space. ICLR Workshop. find
  15. Padó, S. and Lapata, M. (2007). Dependency-based construction of semantic space models. Computational Linguistics, 33(2):161–199. find
  16. Pennington, J., Socher, R., and Manning, C. D. (2014). Glove: Global vectors for word representation. In EMNLP, volume 14, pages 1532–1543. find
  17. Vylomova, E., Cohn, T., He, X., and Haffari, G. (2017). Word representation models for morphologically rich languages in neural machine translation. find
  18. Zahran, M. A., Magooda, A., Mahgoub, A. Y., Raafat, H., Rashwan, M., and Atyia, A. (2015). Word representations in vector space and their applications for arabic. In International Conference on Intelligent Text Processing and Computational Linguistics, pages 430–443. Springer. find
  19. Zou, W. Y., Socher, R., Cer, D., and Manning, C. D. (2013). Bilingual word embeddings for phrase-based machine translation. In Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing, pages 1393–1398. find

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/lingbulk-2020-mersha-morphology-rich-alphasyllabary-embeddings/