Identifying Semitic Roots: Machine Learning with Linguistic Constraints

What it is about

Semitic words are built from a root (usually three consonants) and a pattern, and the root is hard to recover from unpointed text because of weak paradigms, missing vowels and attached prefixes. Earlier systems depend on large hand-built lexicons. The authors train one SNoW Winnow classifier per root radical on letter-position, letter-bigram, prefix and suffix features, and then combine the three outputs with a scoring function that uses limited linguistic knowledge: candidate roots must be in a list of 2,152 Hebrew roots, weak radicals may be deleted, and edit distance to the surface form counts. The data is a hand-tagged newspaper corpus of 5,242 word types covering 1,043 roots. Simple combinations score an F-score of about 45 to 59; restricting the radical sequence and then re-ranking with the linguistic score gives the best results, and the conclusion reports over 80% for its best setup. On roots unseen in training the test F-score is 65.60. Regular roots reach about 95%, while weak paradigms and a 200-word "hard" set (missing radicals, metathesis) are much lower. Human subjects score F 81.86.

Ideas

• I1. A deterministic root guesser for macula dict, which has no roots yet: per-radical rule tables with the weak-paradigm rules (a first radical n or ʾ may vanish, a second w or y, a third y, a doubled second and third) generate candidates, a root list built from the lemmas filters them, and edit distance ranks them. This uses the lexicon instead of a classifier.
• I2. Re-rank, do not only filter: the paper shows that using the root list as a filter alone gives a lower F-score than the full scoring. For Mova, let the scorer order the candidates and write the reason for each candidate's rank.
• I3. Keep their split of results by paradigm (regular, weak, mixed) and their 200-word hard set (a missing radical or metathesis) as a test of the root step in macula dict.

For Mova

I1 and I2 go into macula dict once it has roots, and the weak-paradigm rules are data for The weak verbs. I3 is a test set for tools/macula dict. The paper's data is Modern Hebrew newspaper text, so the numbers do not transfer; the Ancient Hebrew test has to come from MACULA lemmas.

Cite

The paper itself, not this card. BibTeX from the library's record of the work. DOI 10.1162/coli.2008.07-002-r1-06-30

@article{daya2008identifying,
  title = {{Identifying Semitic Roots: Machine Learning with Linguistic Constraints}},
  author = {Daya, Ezra and Roth, Dan and Wintner, Shuly},
  journal = {Computational Linguistics},
  year = {2008},
  doi = {10.1162/coli.2008.07-002-r1-06-30},
  url = {https://doi.org/10.1162/coli.2008.07-002-r1-06-30}
}

References

31 references, 4 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Abu-Salem, Hani, Mahmoud Al-Omari, and Martha W. Evens. 1999. Stemming methodologies over individual query words for an Arabic information retrieval system. Journal of the American Society for Information Science, 50(6):524–529. find
2. Al-Kharashi, Ibrahim A. and Martha W. Evens. 1994. Comparing words, stems, and roots as index terms in an Arabic information retrieval system. Journal of the American Society for Information Science, 45(8):548–560. find
3. Banko, Michele and Eric Brill. 2001. Scaling to very very large corpora for natural language disambiguation. In Proceedings of the 39th Annual Meeting of the Association for Computational Linguistics, pages 26–33, Morristown, NJ. find
4. Beesley, Kenneth R. 1998a. Arabic morphological analysis on the internet. In Proceedings of the 6th International Conference and Exhibition on Multi-lingual Computing, Cambridge, UK. find
5. Beesley, Kenneth R. 1998b. Arabic morphology using only finite-state operations. Proceedings of the Workshop on Computational Approaches to Semitic Languages, pages 50–57, Montreal, Quebec. find our card
6. Buckwalter, Tim. 2002. Buckwalter Arabic morphological analyzer. Linguistic Data Consortium (LDC) catalog number LDC2002L49 and ISBN 1-58563-257-0. find
7. Carlson, Andrew J., Chad M. Cumby, Jeff L. Rosen, and Dan Roth. 1999. The SNoW learning architecture. Technical Report UIUCDCS-R-99-2101, UIUC Computer Science Department. find
8. Choueka, Yaacov. 1990. MLIM—A system for full, exact, on-line grammatical analysis of Modern Hebrew. In Proceedings of the Annual Conference on Computers in Education, page 63, Tel Aviv. [In Hebrew.] find
9. Darwish, Kareem. 2002. Building a shallow Arabic morphological analyzer in one day. In Computational Approaches to Semitic Languages, an ACL’02 Workshop, pages 47–54, Philadelphia, PA. find
10. Darwish, Kareem and Douglas W. Oard. 2002. Term selection for searching printed Arabic. In SIGIR ’02: Proceedings of the 25th Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, pages 261–268, New York, NY. find
11. Daya, Ezra, Dan Roth, and Shuly Wintner. 2004. Learning Hebrew roots: Machine learning with linguistic constraints. In Proceedings of EMNLP’04, pages 357–364, Barcelona, Spain. find our card
12. Daya, Ezra, Dan Roth, and Shuly Wintner. 2007. Learning to identify Semitic roots. In Abdelhadi Soudi, Guenter Neumann, and Antal van den Bosch, editors, Arabic Computational Morphology: Knowledge-based and Empirical Methods, volume 38 of Text, Speech and Language Technology. Springer, New York, pages 143–158. find
13. Even-Shoshan, Abraham. 1993. HaMillon HaXadash (The New Dictionary). Kiryat Sefer, Jerusalem. In Hebrew. find
14. Even-Zohar, Yair and Dan Roth. 2001. A sequential model for multi class classification. In EMNLP-2001, the SIGDAT Conference on Empirical Methods in Natural Language Processing, pages 10–19, Pittsburgh, PA. find
15. Florian, Radu. 2002. Named entity recognition as a house of cards: Classifier stacking. In Proceedings of CoNLL-2002, pages 175–178, Taiwan. find
16. Golding, Andrew R. and Dan Roth. 1999. A Winnow based approach to context-sensitive spelling correction. Machine Learning, 34(1–3):107–130. find
17. Habash, Nizar and Owen Rambow. 2005. Arabic tokenization, part-of-speech tagging and morphological disambiguation in one fell swoop. In Proceedings of the 43rd Annual Meeting of the Association for Computational Linguistics (ACL’05), pages 573–580, Ann Arbor, MI. find our card
18. Kruskal, Joseph. 1999. An overview of sequence comparison. In David Sankoff and Joseph Kruskal, editors, Time Warps, String Edits and Macromolecules: The Theory and Practice of Sequence Comparison. CSLI Publications, Stanford, CA, pages 1–44. find
19. Larkey, Leah S., Lisa Ballesteros, and Margaret E. Connell. 2002. Improving stemming for Arabic information retrieval: Light stemming and co-occurrence analysis. In SIGIR ’02: Proceedings of the 25th Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, pages 275–282, New York, NY. find
20. Levenshtein, Vladimir I. 1965. Binary codes capable of correcting deletions, insertions and reversals. Doklady Akademii Nauk SSSR, 163(4):845–848. find
21. McCarthy, John J. 1981. A prosodic theory of nonconcatenative morphology. Linguistic Inquiry, 12(3):373–418. find
22. Ornan, Uzzi. 2003. The Final Word. University of Haifa Press, Haifa, Israel. [In Hebrew.] find
23. Owens, Jonathan. 1997. The Arabic grammatical tradition. In Robert Hetzron, editor, The Semitic Languages. Routledge, London and New York, chapter 3, pages 46–58. find
24. Punyakanok, Vasin and Dan Roth. 2001. The use of classifiers in sequential inference. In NIPS-13; The 2000 Conference on Advances in Neural Information Processing Systems 13, pages 995–1001, Denver, CO. find
25. Punyakanok, Vasin, Dan Roth, and Wen-Tau Yih. 2005. The necessity of syntactic parsing for semantic role labeling. In Proceedings of IJCAI 2005, pages 1117–1123, Edinburgh. find
26. Roth, Dan. 1998. Learning to resolve natural language ambiguities: A unified approach. In Proceedings of AAAI-98 and IAAI-98, pages 806–813, Madison, WI. find
27. Schütze, Hinrich and Yoram Singer. 1994. Part-of-speech tagging using a variable memory Markov model. In Proceedings of the 32nd Annual Meeting of the Association for Computational Linguistics, pages 181–187, Las Cruses, NM. find
28. Shimron, Joseph, editor. 2003. Language Processing and Acquisition in Languages of Semitic, Root-Based, Morphology. Number 28 in Language Acquisition and Language Disorders. John Benjamins, Amsterdam. find
29. Tjong Kim Sang, Erik F. and Fien De Meulder. 2003. Introduction to the CoNLL-2003 shared task: Language-independent named entity recognition. In Proceedings of CoNLL-2003, pages 142–147, Edmonton, Canada. find
30. Yona, Shlomo and Shuly Wintner. 2008. A finite-state morphological grammar of Hebrew. Natural Language Engineering, 14(2):173–190. find our card
31. Zdaqa, Yizxaq. 1974. Luxot HaPoal [The Verb Tables]. Kiryath Sepher, Jerusalem. [In Hebrew.] find

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/cl-2008-daya-identifying-semitic-roots-machine-learning/