Universal Dependencies Parsing for Colloquial Singaporean English

What it is about

The authors ask why an English parser fails on Singlish, an English-based creole with imported Hokkien, Malay and Cantonese words, topic-first sentences, copula deletion and clause-final particles such as "lah". They annotate a 1,200-sentence Singlish treebank in UD English style, chosen from forum posts that an English language model finds unusually unlikely, and report 93.44 UAS and 89.63 LAS inter-annotator agreement on 100 double-annotated sentences. A biaffine parser trained on the Singlish data alone reaches UAS 79.29. Neural stacking on top of an English model trained on UD English EWT raises it to UAS 84.47 and LAS 77.76 on the 150-sentence test set, a 25.01% relative error reduction, and it also lifts POS accuracy to 89.50%. Embeddings trained on the Singlish corpus help more than GloVe embeddings, which the authors read as a lexical-distribution difference.

Ideas

• I1. en/src/ptag.rs and en/src/ud.rs: tag the clause-final particles (lah, leh, lor, meh, hor, ah) as INTJ and attach them as discourse to the clause head; measure UPOS and relation accuracy on a colloquial dev slice that we build from open forum text.
• I2. en/src/parse.rs: add a feature from a second, small English model trained on colloquial text, as a perceptron analogue of the paper's neural stacking; measure LAS on the same colloquial dev slice, with and without the feature, run twice for determinism.
• I3. en/src/morph.rs with en/src/tok.rs: a dictionary of imported words with their UPOS (the paper's Table A4 gives a starting list), consulted before the suffix rules; measure UPOS accuracy on tokens that are outside the EWT vocabulary.
• I4. tools/treeagree gates: copula-deleted sentences have no cop arc by design, so check that the gates do not reject them; measure gate violations per 1k colloquial silver trees with treeagree check.

For Mova

Touches languages/en/08-national-varieties, 19-lexicon-mwe-register, 21-clause-structure and 38-dialogue-and-discourse-markers, and the en crate (ptag.rs, morph.rs, parse.rs) and tools/treeagree. Measure first the Mova LAS on a colloquial slice built from open text; the paper's 150-sentence test set is a reference only. Its licence (CC BY 4.0) allows reuse, but the first step is to see whether the treebank is on disk.

Topics: Universal Dependencies

Cite

The paper itself, not this card. BibTeX from the library's record of the work. DOI 10.18653/v1/p17-1159 · arXiv 1705.06463

@inproceedings{wang2017universal,
  title = {{Universal Dependencies Parsing for Colloquial Singaporean English}},
  author = {Wang, Hongmin and Zhang, Yue and Chan, GuangYong Leonard and Yang, Jie and Chieu, Hai Leong},
  booktitle = {Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2017},
  doi = {10.18653/v1/p17-1159},
  eprint = {1705.06463},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.18653/v1/p17-1159}
}

References

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

1. Waleed Ammar, George Mulcaire, Miguel Ballesteros, Chris Dyer, and Noah Smith. 2016. Many languages, one parser. Transactions of the Association of Computational Linguistics 4:431–444. http://aclweb.org/anthology/Q16-1031. ↗
2. Daniel Andor, Chris Alberti, David Weiss, Aliaksei Severyn, Alessandro Presta, Kuzman Ganchev, Slav Petrov, and Michael Collins. 2016. Globally normalized transition-based neural networks. In Proceedings of the ACL 2016. Association for Computational Linguistics, pages 2442–2452. https://doi.org/10.18653/v1/P16-1231. ↗
3. John C. Duchi, Elad Hazan, and Yoram Singer. 2011. Adaptive subgradient methods for online learning and stochastic optimization. Journal of Machine Learning Research 12:2121–2159. http://dl.acm.org/citation.cfm?id=2021068. ↗
4. Long Duong, Trevor Cohn, Steven Bird, and Paul Cook. 2015. A neural network model for low-resource universal dependency parsing. In Proceedings of the EMNLP 2015. Association for Computational Linguistics, pages 339–348. https://doi.org/10.18653/v1/D15-1040. ↗
5. Dzmitry Bahdanau, Kyunghyun Cho, and Yoshua Bengio. 2014. Neural machine translation by jointly learning to align and translate. arXiv preprint abs/1409.0473. http://arxiv.org/abs/1409.0473. ↗
6. Chris Dyer, Miguel Ballesteros, Wang Ling, Austin Matthews, and A. Noah Smith. 2015. Transition-based dependency parsing with stack long short-term memory. In Proceedings of the ACL-IJCNLP 2015. Association for Computational Linguistics, pages 334–343. https://doi.org/10.3115/v1/P15-1033. ↗ our card
7. Miguel Ballesteros, Chris Dyer, and A. Noah Smith. 2015. Improved transition-based parsing by modeling characters instead of words with lstms. In Proceedings of the EMNLP 2015. Association for Computational Linguistics, pages 349–359. https://doi.org/10.18653/v1/D15-1041. ↗
8. Kuzman Ganchev, Jennifer Gillenwater, and Ben Taskar. 2009. Dependency grammar induction via bitext projection constraints. In Proceedings of the ACL-IJCNLP 2009. Association for Computational Linguistics, pages 369–377. http://aclweb.org/anthology/P09-1042. ↗
9. Ann Bies, Justin Mott, Colin Warner, and Seth Kulick. 2012. English web treebank ldc2012t13 . find
10. Douwe Gelling, Trevor Cohn, Phil Blunsom, and Joao Graca. 2012. The pascal challenge on grammar induction. In Proceedings of the NAACL-HLT Workshop on the Induction of Linguistic Structure. Association for Computational Linguistics, pages 64–80. http://www.aclweb.org/anthology/W12-1909. ↗
11. Danqi Chen and Christopher Manning. 2014. A fast and accurate dependency parser using neural networks. In Proceedings of the EMNLP 2014. Association for Computational Linguistics, pages 740–750. https://doi.org/10.3115/v1/D14-1082. ↗ our card
12. Hongshen Chen, Yue Zhang, and Qun Liu. 2016. Neural network for heterogeneous annotations. In Proceedings of the EMNLP 2016. Association for Computational Linguistics, pages 731–741. http://aclweb.org/anthology/D16-1070. ↗
13. Shay Cohen and A. Noah Smith. 2009. Shared logistic normal distributions for soft parameter tying in unsupervised grammar induction. In Proceedings of the NAACL-HLT 2009. Association for Computational Linguistics, pages 74–82. http://aclweb.org/anthology/N09-1009. ↗
14. Ronan Collobert, Jason Weston, Léon Bottou, Michael Karlen, Koray Kavukcuoglu, and Pavel Kuksa. 2011. Natural language processing (almost) from scratch. Journal of Machine Learning Research 12:2493–2537. http://dl.acm.org/citation.cfm?id=2078186. ↗
15. Timothy Dozat and Christopher D. Manning. 2017. Deep biaffine attention for neural dependency parsing. In International Conference on Learning Representations 2017. volume abs/1611.01734. http://arxiv.org/abs/1611.01734. ↗ our card
16. Jennifer Gillenwater, Kuzman Ganchev, João Graça, Fernando Pereira, and Ben Taskar. 2010. Sparsity in dependency grammar induction. In Proceedings of the ACL 2010 (Short Papers). Association for Computational Linguistics, pages 194–199. http://www.aclweb.org/anthology/P10-2036. ↗
17. Alex Graves and Jürgen Schmidhuber. 2005. Frame-wise phoneme classification with bidirectional lstm and other neural network architectures. Neural Networks 18(5):602–610. find
18. K. Greff, R. K. Srivastava, J. Koutnk, B. R. Ste-unebrink, and J. Schmidhuber. 2016. Lstm: A search space odyssey. IEEE Transactions on Neural Networks and Learning Systems PP(99):1–11. https://doi.org/10.1109/TNNLS.2016.2582924. ↗
19. Jiang Guo, Wanxiang Che, David Yarowsky, Haifeng Wang, and Ting Liu. 2015. Cross-lingual dependency parsing based on distributed representations. In Proceedings of the ACL-IJCNLP 2015. Association for Computational Linguistics, pages 1234– 1244. https://doi.org/10.3115/v1/P15-1119. ↗
20. Shinichi Harada. 2009. The roles of singapore standard english and singlish. Information Research 40:70– 82. find
21. Kenneth Heafield, Ivan Pouzyrevsky, H. Jonathan Clark, and Philipp Koehn. 2013. Scalable modified kneser-ney language model estimation. In Proceedings of the ACL 2013 (Short Papers). Association for Computational Linguistics, pages 690–696. http://aclweb.org/anthology/P13-2121. ↗
22. Héctor Martı́nez Alonso, Željko Agić, Barbara Plank, and Anders Søgaard. 2017. Parsing universal dependencies without training. In Proceedings of the EACL 2017. Association for Computational Linguistics, pages 230–240. http://www.aclweb.org/anthology/E17-1022. ↗
23. Sanjika Hewavitharana, Nguyen Bach, Qin Gao, Vamshi Ambati, and Stephan Vogel. 2011. Proceedings of the Sixth Workshop on Statistical Machine Translation, Association for Computational Linguistics, chapter CMU Haitian Creole-English Translation System for WMT 2011, pages 386–392. http://aclweb.org/anthology/W11-2146. ↗
24. Ryan McDonald, Slav Petrov, and Keith Hall. 2011. Multi-source transfer of delexicalized dependency parsers. In Proceedings of the EMNLP 2011. Association for Computational Linguistics, pages 62–72. http://aclweb.org/anthology/D11-1006. ↗
25. Chang Hu, Philip Resnik, Yakov Kronrod, Vladimir Eidelman, Olivia Buzek, and B. Benjamin Bed-erson. 2011. Proceedings of the Sixth Workshop on Statistical Machine Translation, Association for Computational Linguistics, chapter The Value of Monolingual Crowdsourcing in a Real-World Translation Scenario: Simulation using Haitian Creole Emergency SMS Messages, pages 399–404. http://aclweb.org/anthology/W11-2148. ↗
26. Zhiheng Huang, Wei Xu, and Kai Yu. 2015. Bidirectional LSTM-CRF models for sequence tagging. arXiv preprint abs/1508.01991. http://arxiv.org/abs/1508.01991. ↗
27. Rebecca Hwa, Philip Resnik, Amy Weinberg, Clara Cabezas, and Okan Kolak. 2005. Bootstrapping parsers via syntactic projection across parallel texts. Natural Language Engineering 11(3):311– 325. https://doi.org/10.1017/S1351324905003840. ↗
28. Eliyahu Kiperwasser and Yoav Goldberg. 2016. Simple and accurate dependency parsing using bidirectional lstm feature representations. Transactions of the Association of Computational Linguistics 4:313– 327. http://aclweb.org/anthology/Q16-1023. ↗ our card
29. Karen Lahousse and Béatrice Lamiroy. 2012. Word order in french, spanish and italian: A grammaticalization account. Folia Linguistica 46(2):387–415. find
30. Jakob R. E. Leimgruber. 2009. Modelling variation in Singapore English. Ph.D. thesis, Oxford University. find
31. Jakob R. E. Leimgruber. 2011. Singapore english. Language and Linguistics Compass 5(1):47–62. https://doi.org/10.1111/j.1749-818X.2010.00262.x. ↗
32. Jiwei Li, Thang Luong, Dan Jurafsky, and Eduard Hovy. 2015. When are tree structures necessary for deep learning of representations? In Proceedings of the EMNLP 2015. Association for Computational Linguistics, pages 2304–2314. https://doi.org/10.18653/v1/D15-1278. ↗
33. Lisa Lim. 2007. Mergers and acquisitions: on the ages and origins of singapore english particles. World Englishes 26(4):446–473. find
34. Ho Mian-Lian and John T. Platt. 1993. Dynamics of a contact continuum: Singaporean English. Oxford University Press, USA. find
35. Makoto Miwa and Mohit Bansal. 2016. End-to-end relation extraction using lstms on sequences and tree structures. In Proceedings of the ACL 2016. Association for Computational Linguistics, pages 1105– 1116. https://doi.org/10.18653/v1/P16-1105. ↗
36. Tahira Naseem, Regina Barzilay, and Amir Globerson. 2012. Selective sharing for multilingual dependency parsing. In Proceedings of the ACL 2012. Association for Computational Linguistics, pages 629–637. http://aclweb.org/anthology/P12-1066. ↗
37. Tahira Naseem, Harr Chen, Regina Barzilay, and Mark Johnson. 2010. Using universal linguistic knowledge to guide grammar induction. In Proceedings of the EMNLP 2010. Association for Computational Linguistics, Cambridge, MA, pages 1234– 1244. http://www.aclweb.org/anthology/D10-1120. ↗
38. Paroo Nihilani. 1992. The international computerized corpus of english. Words in a cultural context. Singapore: UniPress pages 84–88. find
39. Joakim Nivre, Marie-Catherine de Marneffe, Filip Ginter, Yoav Goldberg, Jan Hajic, Christopher D. Manning, Ryan McDonald, Slav Petrov, Sampo Pyysalo, Natalia Silveira, Reut Tsarfaty, and Daniel Zeman. 2016. Universal dependencies v1: A multilingual treebank collection. In Proceedings of the LREC 2016. European Language Resources Association. find our card
40. Joakim Nivre, Johan Hall, Sandra Kübler, Ryan Mc-Donald, Jens Nilsson, Sebastian Riedel, and Deniz Yuret. 2007. find
41. The CoNLL 2007 shared task on dependency parsing. In Proceedings of the CoNLL Shared Task Session of EMNLP-CoNLL 2007. Association for Computational Linguistics, pages 915– 932. http://www.aclweb.org/anthology/D/D07/D07-1096. ↗
42. Vincent B Y Ooi. 1997. Analysing the Singapore ICE corpus for lexicographic evidence. ENGLISH LANGUAGE & LITERATURE. http://scholarbank.nus.edu.sg/handle/10635/133118. ↗
43. Jeffrey Pennington, Richard Socher, and Christopher Manning. 2014. Glove: Global vectors for word representation. In Proceedings of the EMNLP 2014. Association for Computational Linguistics, pages 1532–1543. https://doi.org/10.3115/v1/D14-1162. ↗
44. Barbara Plank, Anders Søgaard, and Yoav Goldberg. 2016. Multilingual part-of-speech tagging with bidirectional long short-term memory models and auxiliary loss. In Proceedings of the ACL 2016 (Short Papers). Association for Computational Linguistics, pages 412–418. https://doi.org/10.18653/v1/P16-2067. ↗
45. A Chun-Wei Seah, Hai Leong Chieu, Kian Ming Adam Chai, Loo-Nin Teow, and Lee Wei Yeong. 2015. Troll detection by domain-adapting sentiment analysis. In 18th International Conference on Information Fusion (Fusion) 2015. IEEE, pages 792–799. find
46. Richard Socher, Alex Perelygin, Jean Wu, Jason Chuang, D. Christopher Manning, Andrew Ng, and Christopher Potts. 2013. Recursive deep models for semantic compositionality over a sentiment treebank. In Proceedings of the EMNLP 2013. Association for Computational Linguistics, pages 1631– 1642. http://aclweb.org/anthology/D13-1170. ↗
47. Anders Søgaard. 2012a. Two baselines for unsupervised dependency parsing. In Proceedings of the NAACL-HLT Workshop on the Induction of Linguistic Structure. Association for Computational Linguistics, pages 81–83. http://www.aclweb.org/anthology/W12-1910. ↗
48. Anders Søgaard. 2012b. Unsupervised dependency parsing without training. Natural Language Engineering 18(2):187203. https://doi.org/10.1017/S1351324912000022. ↗
49. Oscar Täckström, Ryan McDonald, and Jakob Uszkoreit. 2012. Cross-lingual word clusters for direct transfer of linguistic structure. In Proceedings of the NAACL-HLT 2012. Association for Computational Linguistics, pages 477–487. http://aclweb.org/anthology/N12-1052. ↗
50. David Weiss, Chris Alberti, Michael Collins, and Slav Petrov. 2015. Structured training for neural network transition-based parsing. In Proceedings of the ACL-IJCNLP 2015. Association for Computational Linguistics, pages 323–333. https://doi.org/10.3115/v1/P15-1032. ↗
51. Yuan Zhang and Regina Barzilay. 2015. Hierarchical low-rank tensors for multilingual transfer parsing. In Proceedings of the EMNLP 2015. Association for Computational Linguistics, pages 1857– 1867. https://doi.org/10.18653/v1/D15-1213. ↗
52. Yuan Zhang and David Weiss. 2016. Stack-propagation: Improved representation learning for syntax. In Proceedings of the 54th ACL. Association for Computational Linguistics, pages 1557– 1566. https://doi.org/10.18653/v1/P16-1147. ↗
53. Hao Zhou, Yue Zhang, Shujian Huang, and Jiajun Chen. 2015. A neural probabilistic structured-prediction model for transition-based dependency parsing. In Proceedings of the ACL-IJCNLP 2015. Association for Computational Linguistics, pages 1213–1222. https://doi.org/10.3115/v1/P15-1117. Statistics of Singlish Dependency Treebank ADJ ADP ADV AUX CONJ DET 782 490 941 429 167 387 POS Tags INTJ 556 NOUN 1779 NUM 153 PART 355 PRON 682 PROPN 810 PUNCT SCONJ SYM VERB X 1604 126 11 1704 10 Table A1: Statistics of POS tags ↗

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/ud-2017-wang-universal-dependencies-parsing-colloquial-singaporean/