Elamur Research⌕
☰
home / papers / 2020 / The ADAPT Enhanced Dependency Parser at the IWPT 2020 Shared Task

The ADAPT Enhanced Dependency Parser at the IWPT 2020 Shared Task

Summary

The paper describes the ADAPT team's system for the IWPT 2020 shared task. The task was to obtain enhanced Universal Dependencies (enhanced UD) from raw text for 17 languages, that is, graphs in which a word can have multiple heads. The system works as a pipeline. UDPipe splits the text into sentences and tokens. Next, ensembles of UDPipe-future build the basic tree and simultaneously predict lemmas, POS tags and morphology. These models are augmented with multi-treebank training and external ELMo and FastText embeddings. The enhanced graph is built in two ways: with simple rules on top of the basic tree, or with a graph-based semantic parser following Dozat and Manning. This parser decides separately for each pair of words whether there is an edge between them. The semantic parser proved better than the rules for almost all treebanks. However, the system did not guarantee graph connectivity. The organizers' tool, used to fix this at the last moment, sharply lowered the score: eighth place with an ELAS of 67.23. The team's own fragment-connection algorithm raised the result to 79.53, which would have corresponded to fourth place. The paper's conclusion: the semantic parsing approach works well for enhanced UD, but graph connectivity must be ensured within the system itself.

Key points

  • System components: UDPipe with UD v2.5 models, UDPipe-future with multi-treebank models (treebank ID embedding of size 12), ensembles of 3, 5 or 7 models. ELMo embeddings were taken from ElmoForManyLangs for 15 languages; FastText was trained on the CoNLL'17 raw texts. No external embeddings were used for Lithuanian and Tamil.
  • The semantic parser is based on a biaffine architecture. For edges it uses a sigmoid loss with a threshold of 0.5, for labels softmax. Its input consists of BERT embeddings and character-level BiLSTM representations. The parser predicts labels containing lemmas and cases as atomic units.
  • On the dev sets, the semantic parser outperformed the rules everywhere except two treebanks (nl_lassysmall and et_edt). For uk_iu, ELAS F1 was 82.41 versus 76.88.
  • The official quick-fix attached all unreachable nodes to the root, which drastically reduced accuracy. On the Czech-CAC dev set, ELAS accuracy dropped from 87.37 to 62.26.
  • On the test set, the average across treebanks rose from 67.49 to 79.76 thanks to the team's own fragment connector, and to 80.15 after a rerun with bugs fixed. For Ukrainian-IU, the result rose from 66.43 to 79.81, and then to 82.92. The code is available on GitHub (jbrry/Enhanced-UD-Parsing).

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: Linguistics & languages

Topics: Dependency parsing · Treebanks

Cited by

4 papers of our library cite it (of the 140673 whose bibliographies we read); 2 of them have a page here, 2 not yet.

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2020.iwpt-1.24 · arXiv 2009.01712

@inproceedings{barry2020adapt,
  title = {{The ADAPT Enhanced Dependency Parser at the IWPT 2020 Shared Task}},
  author = {Barry, James and Wagner, Joachim and Foster, Jennifer},
  booktitle = {Proceedings of the 16th International Conference on Parsing Technologies and the IWPT 2020 Shared Task on Parsing into Enhanced Universal Dependencies},
  year = {2020},
  doi = {10.18653/v1/2020.iwpt-1.24},
  eprint = {2009.01712},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.18653/v1/2020.iwpt-1.24}
}

References

25 references, 10 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Mikhail Arkhipov, Maria Trofimova, Yuri Kuratov, and Alexey Sorokin. 2019. Tuning multilingual transformers for language-specific named entity recognition. In Proceedings of the 7th Workshop on Balto-Slavic Natural Language Processing, pages 89–93, Florence, Italy. Association for Computational Ling… find our card
  2. Giuseppe Attardi and Felice Dell’Orletta. 2009. Reverse revision and linear tree combination for dependency parsing. In Proceedings of Human Language Technologies: The 2009 Annual Conference of the North American Chapter of the Association for Computational Linguistics, Companion Volume: Short Paper… find
  3. Piotr Bojanowski, Edouard Grave, Armand Joulin, and Tomas Mikolov. 2017. Enriching word vectors with subword information. Transactions of the Association for Computational Linguistics, 5:135–146. find
  4. Gosse Bouma, Djamé Seddah, and Daniel Zeman. 2020. Overview of the IWPT 2020 Shared Task on Parsing into Enhanced Universal Dependencies. In Proceedings of the 16th International Conference on Parsing Technologies and the IWPT 2020 Shared Task on Parsing into Enhanced Universal Dependencies, Seattl… find our card
  5. Wanxiang Che, Yijia Liu, Yuxuan Wang, Bo Zheng, and Ting Liu. 2018. Towards better UD parsing: Deep contextualized word embeddings, ensemble, and treebank concatenation. In Proceedings of the CoNLL 2018 Shared Task: Multilingual Parsing from Raw Text to Universal Dependencies, pages 55–64, Brussels,… find
  6. Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. BERT: Pre-training of deep bidirectional transformers for language understanding. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies,… find
  7. Timothy Dozat and Christopher D. Manning. 2017. Deep biaffine attention for neural dependency parsing. In Proceedings of the 5th International Conference on Learning Representations (ICLR 2017). find our card
  8. Timothy Dozat and Christopher D. Manning. 2018. Simpler but more accurate semantic dependency parsing. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 484–490, Melbourne, Australia. Association for Computational Linguistics. find
  9. Matt Gardner, Joel Grus, Mark Neumann, Oyvind Tafjord, Pradeep Dasigi, Nelson F. Liu, Matthew Peters, Michael Schmitz, and Luke Zettlemoyer. 2018. AllenNLP: A deep semantic natural language processing platform. In Proceedings of Workshop for NLP Open Source Software (NLP-OSS), pages 1– 6, Melbourne,… find our card
  10. Johan Hall, Jens Nilsson, Joakim Nivre, Gülşen Eryiǧit, Beáta Megyesi, Mattias Nilsson, and Markus Saers. 2007. Single malt or blended? a study in multilingual parser optimization. In Proceedings of the 2007 Joint Conference on Empirical Methods in Natural Language Processing and Computational N… find
  11. Yuri Kuratov and Mikhail Arkhipov. 2019. Adaptation of deep bidirectional multilingual transformers for Russian language. ArXiv 1905.07213. find
  12. Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, and Veselin Stoyanov. 2019. RoBERTa: A robustly optimized BERT pretraining approach. ArXiv 1907.11692. find
  13. Louis Martin, Benjamin Muller, Pedro Javier Ortiz Suárez, Yoann Dupont, Laurent Romary, Éric Villemonte de la Clergerie, Djamé Seddah, and Benoı̂t Sagot. 2020. CamemBERT: a tasty French language model. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics (ACL… find our card
  14. Matthew Peters, Mark Neumann, Mohit Iyyer, Matt Gardner, Christopher Clark, Kenton Lee, and Luke Zettlemoyer. 2018. Deep contextualized word representations. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technolog… find
  15. Peng Qi, Timothy Dozat, Yuhao Zhang, and Christopher D. Manning. 2018. Universal dependency parsing from scratch. In Proceedings of the CoNLL 2018 Shared Task: Multilingual Parsing from Raw Text to Universal Dependencies, pages 160–170, Brussels, Belgium. Association for Computational Linguistics. find
  16. Ali Safaya, Moutasem Abdullatif, and Deniz Yuret. 2020. Kuisail at semeval-2020 task 12: BERT-CNN for offensive speech identification in social media. In Proceedings of the International Workshop on Semantic Evaluation (SemEval). To appear. find
  17. Milan Straka. 2018. UDPipe 2.0 prototype at CoNLL 2018 UD shared task. In Proceedings of the CoNLL 2018 Shared Task: Multilingual Parsing from Raw Text to Universal Dependencies, pages 197–207, Brussels, Belgium. Association for Computational Linguistics. find our card
  18. Milan Straka and Jana Straková. 2017. Tokenizing, POS tagging, lemmatizing and parsing UD 2.0 with UDPipe. In Proceedings of the CoNLL 2017 Shared Task: Multilingual Parsing from Raw Text to Universal Dependencies, pages 88–99, Vancouver, Canada. Association for Computational Linguistics. find our card
  19. Milan Straka and Jana Straková. 2019. Universal dependencies 2.5 models for UDPipe (2019-12-06). LINDAT/CLARIAH-CZ digital library at the Institute of Formal and Applied Linguistics (ÚFAL), Faculty of Mathematics and Physics, Charles University. find
  20. Milan Straka, Jana Straková, and Jan Hajič. 2019. Evaluating contextualized embeddings on 54 languages in POS tagging, lemmatization and dependency parsing. ArXiv 1908.07448. find
  21. Sara Stymne, Miryam de Lhoneux, Aaron Smith, and Joakim Nivre. 2018. Parser training with heterogeneous treebanks. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 619–625. Association for Computational Linguistics. find
  22. Antti Virtanen, Jenna Kanerva, Rami Ilo, Jouni Luoma, Juhani Luotolahti, Tapio Salakoski, Filip Ginter, and Sampo Pyysalo. 2019. Multilingual is not enough: BERT for Finnish. ArXiv 1912.07076. find our card
  23. Wietse de Vries, Andreas van Cranenburgh, Arianna Bisazza, Tommaso Caselli, Gertjan van Noord, and Malvina Nissim. 2019. BERTje: A Dutch BERT model. Arxiv 1912.09582. find our card
  24. Joachim Wagner, James Barry, and Jennifer Foster. 2020. Treebank embedding vectors for out-of-domain dependency parsing. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics (ACL). To appear. Also available as ArXiv 2005.00800. find
  25. Thomas Wolf, Lysandre Debut, Victor Sanh, Julien Chaumond, Clement Delangue, Anthony Moi, Pier-ric Cistac, Tim Rault, Rémi Louf, Morgan Funtow-icz, and Jamie Brew. 2019. HuggingFace’s transformers: State-of-the-art natural language processing. ArXiv 1910.03771. find our card

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/barry-2020-adapt-enhanced-dependency-parser-iwpt/