{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/parsing-tweets-into-universal-dependencies","title":"Parsing Tweets into Universal Dependencies","arxiv_id":"1804.08228","date":"2018-04-23","proceeding":"NAACL 2018 6","authors":["Yijia Liu","Yi Zhu","Wanxiang Che","Bing Qin","Nathan Schneider","Noah A. Smith"],"abstract":"We study the problem of analyzing tweets with Universal Dependencies. We\nextend the UD guidelines to cover special constructions in tweets that affect\ntokenization, part-of-speech tagging, and labeled dependencies. Using the\nextended guidelines, we create a new tweet treebank for English (Tweebank v2)\nthat is four times larger than the (unlabeled) Tweebank v1 introduced by Kong\net al. (2014). We characterize the disagreements between our annotators and\nshow that it is challenging to deliver consistent annotation due to ambiguity\nin understanding and explaining tweets. Nonetheless, using the new treebank, we\nbuild a pipeline system to parse raw tweets into UD. To overcome annotation\nnoise without sacrificing computational efficiency, we propose a new method to\ndistill an ensemble of 20 transition-based parsers into a single one. Our\nparser achieves an improvement of 2.2 in LAS over the un-ensembled baseline and\noutperforms parsers that are state-of-the-art on other treebanks in both\naccuracy and speed.","url_abs":"http://arxiv.org/abs/1804.08228v1","url_pdf":"http://arxiv.org/pdf/1804.08228v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"parsing-tweets-into-universal-dependencies","repo_url":"https://github.com/Oneplus/Tweebank","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[{"task_slug":"computational-efficiency","task_name":"Computational Efficiency"},{"task_slug":"dependency-parsing","task_name":"Dependency Parsing"},{"task_slug":"part-of-speech-tagging","task_name":"Part-Of-Speech Tagging"}],"methods":[],"datasets_introduced":[{"slug":"tweebank","name":"Tweebank","full_name":""}],"methods_introduced":[],"results":[{"leaderboard":"/sota/dependency-parsing-on-tweebank","task":"Dependency Parsing","dataset":"Tweebank","model":"Ensemble (20)","rank_in_archive_order":2,"of":3,"metrics":{"Labelled Attachment Score":"79.4","Unlabeled Attachment Score":"83.4"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1804.08228","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}