Noun Phrase Chunking in Hebrew: Influence of Lexical and Morphological Features

What it is about

The traditional base-NP definition (non-recursive NPs, as in English) gives tiny, one-word chunks on the Hebrew Treebank, so the authors define "Simple NPs" instead. These allow smixut (construct chains) and possessives with "shel", but block embedded relatives, prepositional phrases, verb phrases and most conjunctions. They show that English-style chunkers score poorly on Hebrew (F about 76 to 86). An SVM over word, tag and morphology features reaches F 93.2 on 5,000 sentences (27K chunks) with gold morphology, and F 91.4 with the authors' own analyser, which has about 10% errors. Number and construct features add about 0.75 F with gold morphology and about 3 F with noisy morphology; the gender feature hurts. Conjunctions cause the most errors, followed by split and merge errors.

Ideas

• I1. Use the Simple NP definition as the rule for NP spans in Hebrew: stop a span at PP, relative clause, VP, apposition comma and most conjunctions, but keep smixut and "shel" possessives inside. Write it as a rule over the UD tree, and measure it against the NP spans the trees imply.
• I2. Do not put the gender feature into any Hebrew span or agreement rule without a test: keep number and construct, and check by ablation that gender lowers the score on our data too.
• I3. Use the paper's seven error classes (conjunction, split, merge, short, long, whole chunk, missing token) as the diagnostic report for any Hebrew NP output, so that the errors are counted by kind.

For Mova

I1 goes into 7. Syntax and into the smixut section of Syntax: the clause and the sentence; it is a rule for the NP spans of prag. I2 is a feature choice for Nouns, adjectives, pronouns and numerals, measured by ablation. I3 is a report in the test code for treebanks/he. The paper is 2006 and uses shallow chunking; Mova needs the definition and the ablation, not the SVM.

Cite

The paper itself, not this card. BibTeX from the library's record of the work. DOI 10.3115/1220175.1220262

@inproceedings{goldberg2006noun,
  title = {{Noun Phrase Chunking in Hebrew: Influence of Lexical and Morphological Features}},
  author = {Goldberg, Yoav and Adler, Meni and Elhadad, Michael},
  booktitle = {Proceedings of the 21st International Conference on Computational Linguistics and 44th Annual Meeting of the Association for Computational Linguistics},
  year = {2006},
  doi = {10.3115/1220175.1220262},
  url = {https://doi.org/10.3115/1220175.1220262}
}

References

16 references, 2 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Meni Adler and Michael Elhadad, 2006. Unsupervised Morpheme-based HMM for Hebrew Morphological Disambiguation. In Proc. of COLING/ACL 2006, Sidney. find
2. Steven P. Abney. 1991. Parsing by Chunks. In Robert C. Berwick, Steven P. Abney, and Carol Tenny editors, Principle Based Parsing. Kluwer Academic Publishers. find
3. Erin L. Allwein, Robert E. Schapire, and Yoram Singer. 2000. Reducing Multiclass to Binary: A Unifying Approach for Margin Classifiers. Journal of Machine Learning Research, 1:113-141. find
4. Claire Cardie and David Pierce. 1998. Error-Driven Pruning of Treebank Grammars for Base Noun Phrase Identification. In Proc. of COLING-98, Montréal. find
5. Mona Diab, Kadri Hacioglu, and Daniel Jurafsky. 2004. Automatic Tagging of Arabic Text: From Raw Text to Base Phrase Chunks, In Proc. of HLT/NAACL 2004, Boston. find
6. Nizar Habash and Owen Rambow, 2005. Arabic Tokenization, Part-of-speech Tagging and Morphological Disambiguation in One Fell Swoop. In Proc. of ACL 2005, Ann Arbor. find our card
7. Thorsten Joachims. 1998. Text Categorization with Support Vector Machines: Learning with Many Relevant Features. In Proc. of ECML-98, Chemnitz. find
8. Taku Kudo and Yuji Matsumato. 2000. Use of Support Vector Learning for Chunk Identification. In Proc. of CoNLL-2000 and LLL-2000, Lisbon. find
9. Taku Kudo and Yuji Matsumato. 2003. Fast Methods for Kernel-Based Text Analysis. In Proc. of ACL 2003, Sapporo. find
10. Yael Netzer-Dahan and Michael Elhadad, 1998. Generation of Noun Compounds in Hebrew: Can Syntactic Knowledge be Fully Encapsulated? In Proc. of INLG-98, Ontario. find our card
11. Lance A. Ramshaw and Mitchel P. Marcus. 1995. Text Chunking Using Transformation-based Learning. In Proc. of the 3 rd ACL Workshop on Very Large Corpora. Cambridge. find
12. Khalil Sima’an, Alon Itai, Yoad Winter, Alon Altman and N. Nativ, 2001. Building a Tree-bank of Modern Hebrew Text, in Traitement Automatique des Langues 42(2). find
13. Fei Sha and Fernando Pereira. 2003. Shallow Parsing with Conditional Random Fields. Technical Report CIS TR MS-CIS-02-35, University of Pennsylvania. find
14. Erik F. Tjong Kim Sang and Sabine Buchholz. 2000. Introduction to the CoNLL-2000 Shared Task: Chunking. In Proc. of CoNLL-2000 and LLL-2000, Lisbon. find
15. Vladimir Vapnik. 1995. The Nature of Statistical Learning Theory. Springer Verlag, New York, NY. find
16. Tong Zhang, Fred Damerau and David Johnson. 2002. Text Chunking based on a Generalization of Winnow. Journal of Machine Learning Research, 2: 615-637. find

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/aclbulk-2006-goldberg-noun-phrase-chunking-hebrew-influence/