---
title: "Selective Sharing for Multilingual Dependency — Naseem 2012"
description: "The paper addresses dependency parsing for a language with no treebank, where the available annotated source languages may be typologically distant from…"
authors: ["Tahira Naseem", "Regina Barzilay", "Amir Globerson"]
year: 2012
venue: "Proceedings of the 50th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)"
source: "https://aclanthology.org/P12-1066/"
licence: "CC BY-NC-SA 3.0 (ACL Anthology, before 2016)"
licence_url: "https://creativecommons.org/licenses/by-nc-sa/3.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/naseem-2012-selective-sharing-multilingual-dependency-parsing/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/naseem-2012-selective-sharing-multilingual-dependency-parsing/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2012](https://md.research.elamur.ai/free/en/papers/2012/page/2/) / Selective Sharing for Multilingual Dependency Parsing

# Selective Sharing for Multilingual Dependency Parsing

**Authors:** [Tahira Naseem](https://md.research.elamur.ai/free/en/authors/tahira-naseem/), [Regina Barzilay](https://md.research.elamur.ai/free/en/authors/regina-barzilay/), [Amir Globerson](https://md.research.elamur.ai/free/en/authors/amir-globerson/) · **Year:** 2012 · **Venue:** Proceedings of the 50th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · **Licence:** [CC BY-NC-SA 3.0 (ACL Anthology, before 2016)](https://creativecommons.org/licenses/by-nc-sa/3.0/)

**Links:** [Publisher's page](https://aclanthology.org/P12-1066/) · [PDF](https://aclanthology.org/P12-1066.pdf) · [OpenAlex](https://openalex.org/W2115774663)

**Citations:** 138 citations (openalex-snapshot · 2026-10-14) · [99](#cited-by) in our library, of the 140884 papers whose bibliographies we read

## Summary
The paper addresses dependency parsing for a language with no treebank, where the available annotated source languages may be typologically distant from the target. Earlier transfer methods work well inside rich families such as Indo-European but struggle with unrelated languages. The authors propose a generative model that splits tree generation into two parts. The first part chooses which POS tags become dependents of a head, and it is treated as universal and trained on all source languages. The second part decides left/right placement, and it is shared only between languages that agree on the relevant typological features, with separate decisions for each dependency type. The model can also use latent binary features in place of known typology. It learns from unannotated target text and uses posterior regularization to favor short dependencies. Evaluated on 17 CoNLL languages from 10 families, it beats the multi-source transfer parser of McDonald et al. (2011) by 5.9% on average and by 14.4% on non-Indo-European languages. This makes cross-lingual transfer more useful for languages that resemble none of the training set.

## Key points
- Evaluation covered 17 languages from 10 families, all mapped to a shared coarse POS tagset. Directed dependency accuracy was the metric.
- The full model, with raw target data and observed typology, averages 59.5%. Baselines score 53.6% (Transfer, McDonald et al. 2011) and 50.6% (Mixture, Cohen et al. 2011).
- Ablations:
  - Removing typology without adding latent features costs 2.6%, with especially large losses for Turkish.
  - Latent features recover 1.4% of that loss.
  - Dropping raw target text costs 0.9%.
- Oracle comparisons:
  - Training selection on target annotations gives 61.3%.
  - Training ordering on target annotations gives 63.7%, which shows ordering is the harder part.
  - Full supervision (MLE) reaches 66.8%.
  - A purely unsupervised monolingual variant gets only about 26%.
- Code is released at groups.csail.mit.edu/rbg/code/unidep/.

## Interesting ideas
- **Factor the task by how universal each part is.** The model separates the near-universal "which dependents" question from the language-specific "where do they go" question, so each piece can be shared to a different degree. This pattern fits many cross-lingual tasks where some sub-decisions transfer well and others do not.
- **Share per dependency type, not per whole language.** Two languages might agree on adposition order yet differ on adjective order, so parameters are tied at that finer level. This is a more precise way to pick donor languages than a single language-similarity score.
- **Use typological databases as features.** Known word-order features steer the sharing. When such features are missing, a few learned latent binary features partly substitute for them, which is handy for poorly documented languages.
- **Add unsupervised learning on target text.** Raw target text lets the model pick up patterns that no source language has, such as Arabic's VSO order in the CoNLL data. Mixing transfer with this kind of learning is cheap insurance against coverage gaps in the source set.
- **Use oracle ablations to find the bottleneck.** Swapping in supervised estimates for one component at a time shows which part limits performance. This is a reusable evaluation design for any modular model.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

Fields: [Natural language processing](https://md.research.elamur.ai/free/en/shelves/nlp/)

Topics: [Dependency parsing](https://md.research.elamur.ai/free/en/topics/dependency-parsing/) · [Part-of-speech tagging](https://md.research.elamur.ai/free/en/topics/part-of-speech-tagging/)

## Cited by

Cited by

99 papers of our library cite it (of the 140884 whose bibliographies we read); 10 of them have a page here, 89 not yet.

- [Cross-Lingual Parser Selection for Low-Resource Languages](https://md.research.elamur.ai/free/en/papers/agic-2017-cross-lingual-parser-selection-low/)
- [Universal Dependencies and Quantitative Typological Trends. A Case Study on Word Order](https://md.research.elamur.ai/free/en/papers/alzetta-2018-universal-dependencies-quantitative-typological-trends/)
- [Many Languages, One Parser](https://md.research.elamur.ai/free/en/papers/ammar-2016-many-languages-one-parser/)
- [Cross-lingual Dependency Parsing Based on Distributed Representations](https://md.research.elamur.ai/free/en/papers/guo-2015-cross-lingual-dependency-parsing-based/)
- [How Multilingual is Multilingual BERT?](https://md.research.elamur.ai/free/en/papers/pires-2019-how-multilingual-is-multilingual-bert/)
- [Modeling Language Variation and Universals: A Survey on Typological Linguistics for Natural Language Processing](https://md.research.elamur.ai/free/en/papers/ponti-2019-modeling-language-variation-universals-survey/)
- [Cross-Lingual Alignment of Contextual Word Embeddings, with Applications to Zero-shot Dependency Parsing](https://md.research.elamur.ai/free/en/papers/schuster-2019-cross-lingual-alignment-contextual-word/)
- [Cross-Lingual Dependency Parsing with Universal Dependencies and Predicted PoS Labels](https://md.research.elamur.ai/free/en/papers/tiedemann-2015-cross-lingual-dependency-parsing-universal/)
- [Universal Dependencies Parsing for Colloquial Singaporean English](https://md.research.elamur.ai/free/en/papers/ud-2017-wang-universal-dependencies-parsing-colloquial-singaporean/)
- [75 Languages, 1 Model: Parsing Universal Dependencies Universally](https://md.research.elamur.ai/free/en/papers/ud-2019-kondratyuk-75-languages-1-model-parsing/)

## Cite

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

```bibtex
@inproceedings{naseem2012selective,
  title = {{Selective Sharing for Multilingual Dependency Parsing}},
  author = {Naseem, Tahira and Barzilay, Regina and Globerson, Amir},
  booktitle = {Proceedings of the 50th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2012},
  url = {https://aclanthology.org/P12-1066/}
}
```

- [← previous Resolving Complex Cases of Definite Pronouns: The Winograd Schema Challenge](https://md.research.elamur.ai/free/en/papers/rahman-2012-resolving-complex-cases-definite-pronouns/)
- [next → Semantic Compositionality through Recursive Matrix-Vector Spaces](https://md.research.elamur.ai/free/en/papers/socher-2012-semantic-compositionality-through-recursive-matrix/)
- [↑ back to papers of 2012](https://md.research.elamur.ai/free/en/papers/2012/page/2/)

## References

24 references, 5 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Taylor Berg-Kirkpatrick and Dan Klein. 2010. Phylogenetic grammar induction. In ACL, pages 1288–1297. [find](https://scholar.google.com/scholar?q=Taylor+Berg%2DKirkpatrick+and+Dan+Klein%2E+2010%2E+Phylogenetic+grammar+induction%2E+In+ACL%2C+pages+1288%E2%80%931297%2E)
2. Sabine Buchholz and Erwin Marsi. 2006. CoNLL-X shared task on multilingual dependency parsing. In Proceedings of CoNLL, pages 149–164. [find](https://scholar.google.com/scholar?q=Sabine+Buchholz+and+Erwin+Marsi%2E+2006%2E+CoNLL%2DX+shared+task+on+multilingual+dependency+parsing%2E+In+Proceedings+of+CoNLL%2C+pages+149%E2%80%93164%2E) [our card](https://md.research.elamur.ai/free/en/papers/parsers-2006-buchholz-conll-x-shared-task-multilingual/)
3. David Burkett and Dan Klein. 2008. Two languages are better than one (for syntactic parsing). In Proceedings of EMNLP, pages 877–886. [find](https://scholar.google.com/scholar?q=David+Burkett+and+Dan+Klein%2E+2008%2E+Two+languages+are+better+than+one+%28for+syntactic+parsing%29%2E+In+Proceedings+of+EMNLP%2C+pages+877%E2%80%93886%2E)
4. Shay B. Cohen, Dipanjan Das, and Noah A. Smith. 2011. Unsupervised structure prediction with non-parallel multilingual guidance. In EMNLP, pages 50–61. [find](https://scholar.google.com/scholar?q=Shay+B%2E+Cohen%2C+Dipanjan+Das%2C+and+Noah+A%2E+Smith%2E+2011%2E+Unsupervised+structure+prediction+with+non%2Dparallel+multilingual+guidance%2E+In+EMNLP%2C+pages+50%E2%80%9361%2E)
5. Bernard Comrie. 1989. Language Universals and Linguistic Typology: Syntax and Morphology. Oxford: Blackwell. [find](https://scholar.google.com/scholar?q=Bernard+Comrie%2E+1989%2E+Language+Universals+and+Linguistic+Typology%3A+Syntax+and+Morphology%2E+Oxford%3A+Blackwell%2E)
6. Jason Eisner and Noah A. Smith. 2010. Favor short dependencies: Parsing with soft and hard constraints on dependency length. In Trends in Parsing Technology: Dependency Parsing, Domain Adaptation, and Deep Parsing, pages 121–150. [find](https://scholar.google.com/scholar?q=Jason+Eisner+and+Noah+A%2E+Smith%2E+2010%2E+Favor+short+dependencies%3A+Parsing+with+soft+and+hard+constraints+on+dependency+length%2E+In+Trends+in+Parsing+Technology%3A+Dependency+Parsing%2C+Domain+Adaptation%2C+and)
7. João Graça, Kuzman Ganchev, and Ben Taskar. 2007. Expectation maximization and posterior constraints. In Advances in NIPS, pages 569–576. [find](https://scholar.google.com/scholar?q=Joa%CC%83o+Grac%CC%A7a%2C+Kuzman+Ganchev%2C+and+Ben+Taskar%2E+2007%2E+Expectation+maximization+and+posterior+constraints%2E+In+Advances+in+NIPS%2C+pages+569%E2%80%93576%2E)
8. Joseph H Greenberg. 1963. Some universals of language with special reference to the order of meaningful elements. In Joseph H Greenberg, editor, Universals of Language, pages 73–113. MIT Press. [find](https://scholar.google.com/scholar?q=Joseph+H+Greenberg%2E+1963%2E+Some+universals+of+language+with+special+reference+to+the+order+of+meaningful+elements%2E+In+Joseph+H+Greenberg%2C+editor%2C+Universals+of+Language%2C+pages+73%E2%80%93113%2E+MIT+Press%2E)
9. Z.S. Harris. 1968. Mathematical structures of language. Wiley. [find](https://scholar.google.com/scholar?q=Z%2ES%2E+Harris%2E+1968%2E+Mathematical+structures+of+language%2E+Wiley%2E)
10. Martin Haspelmath, Matthew S. Dryer, David Gil, and Bernard Comrie, editors. 2005. The World Atlas of Language Structures. Oxford University Press. [find](https://scholar.google.com/scholar?q=Martin+Haspelmath%2C+Matthew+S%2E+Dryer%2C+David+Gil%2C+and+Bernard+Comrie%2C+editors%2E+2005%2E+The+World+Atlas+of+Language+Structures%2E+Oxford+University+Press%2E)
11. R. Hwa, P. Resnik, A. Weinberg, C. Cabezas, and O. Kolak. 2005. Bootstrapping parsers via syntactic projection across parallel texts. Journal of Natural Language Engineering, 11(3):311–325. [find](https://scholar.google.com/scholar?q=R%2E+Hwa%2C+P%2E+Resnik%2C+A%2E+Weinberg%2C+C%2E+Cabezas%2C+and+O%2E+Kolak%2E+2005%2E+Bootstrapping+parsers+via+syntactic+projection+across+parallel+texts%2E+Journal+of+Natural+Language+Engineering%2C+11%283%29%3A311%E2%80%93325%2E)
12. Dan Klein and Christopher Manning. 2004. Corpus-based induction of syntactic structure: Models of dependency and constituency. In Proceedings of ACL, pages 478–485. [find](https://scholar.google.com/scholar?q=Dan+Klein+and+Christopher+Manning%2E+2004%2E+Corpus%2Dbased+induction+of+syntactic+structure%3A+Models+of+dependency+and+constituency%2E+In+Proceedings+of+ACL%2C+pages+478%E2%80%93485%2E)
13. Jonas Kuhn. 2004. Experiments in parallel-text based grammar induction. In Proceedings of the ACL, pages 470–477. [find](https://scholar.google.com/scholar?q=Jonas+Kuhn%2E+2004%2E+Experiments+in+parallel%2Dtext+based+grammar+induction%2E+In+Proceedings+of+the+ACL%2C+pages+470%E2%80%93477%2E)
14. Ryan T. McDonald, Slav Petrov, and Keith Hall. 2011. Multi-source transfer of delexicalized dependency parsers. In EMNLP, pages 62–72. [find](https://scholar.google.com/scholar?q=Ryan+T%2E+McDonald%2C+Slav+Petrov%2C+and+Keith+Hall%2E+2011%2E+Multi%2Dsource+transfer+of+delexicalized+dependency+parsers%2E+In+EMNLP%2C+pages+62%E2%80%9372%2E) [our card](https://md.research.elamur.ai/free/en/papers/mcdonald-2011-multi-source-transfer-delexicalized-dependency/)
15. Tahira Naseem, Harr Chen, Regina Barzilay, and Mark Johnson. 2010. Using universal linguistic knowledge to guide grammar induction. In EMNLP, pages 1234– 1244. [find](https://scholar.google.com/scholar?q=Tahira+Naseem%2C+Harr+Chen%2C+Regina+Barzilay%2C+and+Mark+Johnson%2E+2010%2E+Using+universal+linguistic+knowledge+to+guide+grammar+induction%2E+In+EMNLP%2C+pages+1234%E2%80%93+1244%2E)
16. Joakim Nivre, Johan Hall, Sandra Kübler, Ryan McDonald, Jens Nilsson, Sebastian Riedel, and Deniz Yuret. 2007. [find](https://scholar.google.com/scholar?q=Joakim+Nivre%2C+Johan+Hall%2C+Sandra+Ku%CC%88bler%2C+Ryan+McDonald%2C+Jens+Nilsson%2C+Sebastian+Riedel%2C+and+Deniz+Yuret%2E+2007%2E)
17. The CoNLL 2007 shared task on dependency parsing. In Proceedings of the CoNLL Shared Task Session of EMNLP-CoNLL 2007, pages 915–932. [find](https://scholar.google.com/scholar?q=The+CoNLL+2007+shared+task+on+dependency+parsing%2E+In+Proceedings+of+the+CoNLL+Shared+Task+Session+of+EMNLP%2DCoNLL+2007%2C+pages+915%E2%80%93932%2E) [our card](https://md.research.elamur.ai/free/en/papers/nivre-2007-conll-2007-shared-task-dependency/)
18. Slav Petrov, Dipanjan Das, and Ryan McDonald. 2011. A universal part-of-speech tagset. In ArXiv, April. [find](https://scholar.google.com/scholar?q=Slav+Petrov%2C+Dipanjan+Das%2C+and+Ryan+McDonald%2E+2011%2E+A+universal+part%2Dof%2Dspeech+tagset%2E+In+ArXiv%2C+April%2E)
19. David A. Smith and Noah A. Smith. 2004. Bilingual parsing with factored estimation: Using English to parse Korean. In Proceeding of EMNLP, pages 49– 56. [find](https://scholar.google.com/scholar?q=David+A%2E+Smith+and+Noah+A%2E+Smith%2E+2004%2E+Bilingual+parsing+with+factored+estimation%3A+Using+English+to+parse+Korean%2E+In+Proceeding+of+EMNLP%2C+pages+49%E2%80%93+56%2E)
20. Benjamin Snyder, Tahira Naseem, and Regina Barzilay. 2009. Unsupervised multilingual grammar induction. In Proceedings of ACL/AFNLP, pages 73–81. [find](https://scholar.google.com/scholar?q=Benjamin+Snyder%2C+Tahira+Naseem%2C+and+Regina+Barzilay%2E+2009%2E+Unsupervised+multilingual+grammar+induction%2E+In+Proceedings+of+ACL%2FAFNLP%2C+pages+73%E2%80%9381%2E)
21. Anders Søgaard. 2011. Data point selection for cross-language adaptation of dependency parsers. In ACL (Short Papers), pages 682–686. [find](https://scholar.google.com/scholar?q=Anders+S%C3%B8gaard%2E+2011%2E+Data+point+selection+for+cross%2Dlanguage+adaptation+of+dependency+parsers%2E+In+ACL+%28Short+Papers%29%2C+pages+682%E2%80%93686%2E)
22. Dekai Wu. 1997. Stochastic inversion transduction grammars and bilingual parsing of parallel corpora. Computational Linguistics, 23(3):377–403. [find](https://scholar.google.com/scholar?q=Dekai+Wu%2E+1997%2E+Stochastic+inversion+transduction+grammars+and+bilingual+parsing+of+parallel+corpora%2E+Computational+Linguistics%2C+23%283%29%3A377%E2%80%93403%2E) [our card](https://md.research.elamur.ai/free/en/papers/wu-1997-stochastic-inversion-transduction-grammars-bilingual/)
23. Chenhai Xi and Rebecca Hwa. 2005. A backoff model for bootstrapping resources for non-english languages. In Proceedings of EMNLP, pages 851 – 858. [find](https://scholar.google.com/scholar?q=Chenhai+Xi+and+Rebecca+Hwa%2E+2005%2E+A+backoff+model+for+bootstrapping+resources+for+non%2Denglish+languages%2E+In+Proceedings+of+EMNLP%2C+pages+851+%E2%80%93+858%2E)
24. Daniel Zeman and Philip Resnik. 2008. Cross-language parser adaptation between related languages. In Proceedings of the IJCNLP-08 Workshop on NLP for Less Privileged Languages, pages 35–42, January. [find](https://scholar.google.com/scholar?q=Daniel+Zeman+and+Philip+Resnik%2E+2008%2E+Cross%2Dlanguage+parser+adaptation+between+related+languages%2E+In+Proceedings+of+the+IJCNLP%2D08+Workshop+on+NLP+for+Less+Privileged+Languages%2C+pages+35%E2%80%9342%2C+January%2E) [our card](https://md.research.elamur.ai/free/en/papers/zeman-2008-cross-language-parser-adaptation-between/)