{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,2,24]],"date-time":"2026-02-24T17:14:28Z","timestamp":1771953268492,"version":"3.50.1"},"reference-count":47,"publisher":"MIT Press","issue":"3","license":[{"start":{"date-parts":[[2024,3,12]],"date-time":"2024-03-12T00:00:00Z","timestamp":1710201600000},"content-version":"vor","delay-in-days":71,"URL":"https:\/\/creativecommons.org\/licenses\/by-nc-nd\/4.0\/"}],"content-domain":{"domain":["direct.mit.edu"],"crossmark-restriction":true},"short-container-title":[],"published-print":{"date-parts":[[2024,9,1]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:p>Dialogue-level dependency parsing, despite its growing academic interest, often encounters underperformance issues due to resource shortages. A potential solution to this challenge is data augmentation. In recent years, large language models (LLMs) have demonstrated strong capabilities in generation, which can facilitate data augmentation greatly. In this study, we focus on Chinese dialogue-level dependency parsing, presenting three simple and effective strategies with LLM to augment the original training instances, namely word-level, syntax-level, and discourse-level augmentations, respectively. These strategies enable LLMs to either preserve or modify dependency structures, thereby assuring accuracy while increasing the diversity of instances at different levels. We conduct experiments on the benchmark dataset released by Jiang et al. (2023) to validate our approach. Results show that our method can greatly boost the parsing performance in various settings, particularly in dependencies among elementary discourse units. Lastly, we provide in-depth analysis to show the key points of our data augmentation strategies.<\/jats:p>","DOI":"10.1162\/coli_a_00515","type":"journal-article","created":{"date-parts":[[2024,3,12]],"date-time":"2024-03-12T19:46:57Z","timestamp":1710272817000},"page":"867-891","update-policy":"https:\/\/doi.org\/10.1162\/mitpressjournals.corrections.policy","source":"Crossref","is-referenced-by-count":9,"title":["LLM-Assisted Data Augmentation for Chinese Dialogue-Level Dependency Parsing"],"prefix":"10.1162","volume":"50","author":[{"given":"Meishan","family":"Zhang","sequence":"first","affiliation":[{"name":"Harbin Institute of Technology (Shenzhen), Institute of Computing and Intelligence. mason.zms@gmail.com"}]},{"given":"Gongyao","family":"Jiang","sequence":"additional","affiliation":[{"name":"Tianjin University , School of New Media and Communication. jianggongyao@gmail.com"}]},{"given":"Shuang","family":"Liu","sequence":"additional","affiliation":[{"name":"Tianjin University , College of Intelligence and Computing. shuang.liu@tju.edu.cn"}]},{"given":"Jing","family":"Chen","sequence":"additional","affiliation":[{"name":"Information Center of Ministry of Science and Technology. chenjjing83@sina.com"}]},{"given":"Min","family":"Zhang","sequence":"additional","affiliation":[{"name":"Harbin Institute of Technology (Shenzhen), Institute of Computing and Intelligence. zhangmin2021@hit.edu.cn"}]}],"member":"281","published-online":{"date-parts":[[2024,9,1]]},"reference":[{"key":"2024092014245895000_bib1","first-page":"1","article-title":"GPT-4 technical report","author":"Achiam","year":"2023","journal-title":"arXiv preprint arXiv:2303.08774"},{"key":"2024092014245895000_bib2","doi-asserted-by":"publisher","first-page":"928","DOI":"10.18653\/v1\/D15-1109","article-title":"Discourse parsing for multi-party chat dialogues","volume-title":"Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing","author":"Afantenos","year":"2015"},{"key":"2024092014245895000_bib3","first-page":"2721","article-title":"Discourse structure and dialogue acts in multiparty dialogue: The STAC corpus","volume-title":"10th International Conference on Language Resources and Evaluation (LREC 2016)","author":"Asher","year":"2016"},{"key":"2024092014245895000_bib4","first-page":"1","article-title":"Qwen technical report","author":"Bai","year":"2023","journal-title":"arXiv preprint arXiv:2309.16609"},{"key":"2024092014245895000_bib5","first-page":"1877","article-title":"Language models are few-shot learners","volume":"33","author":"Brown","year":"2020","journal-title":"Advances in Neural Information Processing Systems"},{"key":"2024092014245895000_bib6","doi-asserted-by":"publisher","first-page":"1","DOI":"10.3115\/1118078.1118083","article-title":"Building a discourse-tagged corpus in the framework of Rhetorical Structure Theory","volume-title":"Proceedings of the Second SIGdial Workshop on Discourse and Dialogue","author":"Carlson","year":"2001"},{"key":"2024092014245895000_bib7","article-title":"Chinese Dependency Treebank 1.0 LDC2012T05","author":"Che","year":"2012"},{"issue":"240","key":"2024092014245895000_bib8","first-page":"1","article-title":"PaLM: Scaling language modeling with pathways","volume":"24","author":"Chowdhery","year":"2023","journal-title":"Journal of Machine Learning Research"},{"key":"2024092014245895000_bib9","first-page":"1","article-title":"ELECTRA: Pre-training text encoders as discriminators rather than generators","volume-title":"International Conference on Learning Representations","author":"Clark","year":"2019"},{"key":"2024092014245895000_bib10","doi-asserted-by":"publisher","first-page":"657","DOI":"10.18653\/v1\/2020.findings-emnlp.58","article-title":"Revisiting pre-trained models for Chinese natural language processing","volume-title":"Findings of the Association for Computational Linguistics: EMNLP 2020","author":"Cui","year":"2020"},{"key":"2024092014245895000_bib11","first-page":"1","article-title":"AugGPT: Leveraging chatGPT for text data augmentation","author":"Dai","year":"2023","journal-title":"arXiv preprint arXiv:2302.13007"},{"key":"2024092014245895000_bib12","doi-asserted-by":"publisher","first-page":"1513","DOI":"10.18653\/v1\/D19-1162","article-title":"Dependency parsing for spoken dialog systems","volume-title":"Proceedings of the EMNLP-IJCNLP 2019","author":"Davidson","year":"2019"},{"key":"2024092014245895000_bib13","first-page":"1","article-title":"Deep biaffine attention for neural dependency parsing","volume-title":"International Conference on Learning Representations","author":"Dozat","year":"2016"},{"key":"2024092014245895000_bib14","doi-asserted-by":"publisher","first-page":"968","DOI":"10.18653\/v1\/2021.findings-acl.84","article-title":"A survey of data augmentation approaches for NLP","volume-title":"Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021","author":"Feng","year":"2021"},{"key":"2024092014245895000_bib15","doi-asserted-by":"publisher","first-page":"322","DOI":"10.1109\/TASLP.2022.3224302","article-title":"Curriculum-style fine-grained adaption for unsupervised cross-lingual dependency transfer","volume":"31","author":"Guo","year":"2022","journal-title":"IEEE\/ACM Transactions on Audio, Speech, and Language Processing"},{"key":"2024092014245895000_bib16","doi-asserted-by":"publisher","first-page":"9526","DOI":"10.18653\/v1\/2023.findings-acl.607","article-title":"A pilot study on dialogue-level dependency parsing for Chinese","volume-title":"Findings of the Association for Computational Linguistics: ACL 2023","author":"Jiang","year":"2023"},{"key":"2024092014245895000_bib17","doi-asserted-by":"publisher","first-page":"2706","DOI":"10.18653\/v1\/P18-1252","article-title":"Supervised treebank conversion: Data and approaches","volume-title":"Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","author":"Jiang","year":"2018"},{"issue":"1","key":"2024092014245895000_bib18","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1007\/978-3-031-02131-2","article-title":"Dependency parsing","volume":"1","author":"K\u00fcbler","year":"2009","journal-title":"Synthesis Lectures on Human Language Technologies"},{"key":"2024092014245895000_bib19","first-page":"896","article-title":"Pseudo-label: The simple and efficient semi-supervised learning method for deep neural networks","volume-title":"Workshop on Challenges in Representation Learning, ICML","author":"Lee","year":"2013"},{"key":"2024092014245895000_bib20","doi-asserted-by":"publisher","first-page":"8174","DOI":"10.18653\/v1\/2023.emnlp-main.508","article-title":"LLM-enhanced self-training for cross-domain constituency parsing","volume-title":"Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing","author":"Li","year":"2023"},{"key":"2024092014245895000_bib21","doi-asserted-by":"publisher","first-page":"25","DOI":"10.3115\/v1\/P14-1003","article-title":"Text-level discourse dependency parsing","volume-title":"Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","author":"Li","year":"2014"},{"key":"2024092014245895000_bib22","doi-asserted-by":"publisher","first-page":"2386","DOI":"10.18653\/v1\/P19-1229","article-title":"Semi-supervised domain adaptation for dependency parsing","volume-title":"Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics","author":"Li","year":"2019"},{"key":"2024092014245895000_bib23","doi-asserted-by":"publisher","first-page":"191","DOI":"10.1109\/CCNS50731.2020.00049","article-title":"A survey of text data augmentation","volume-title":"2020 International Conference on Computer Communication and Network Security (CCNS)","author":"Liu","year":"2020"},{"key":"2024092014245895000_bib24","doi-asserted-by":"publisher","first-page":"1","DOI":"10.3115\/1075812.1075835","article-title":"Decoupled weight decay regularization","volume-title":"International Conference on Learning Representations","author":"Loshchilov","year":"2019"},{"key":"2024092014245895000_bib25","doi-asserted-by":"crossref","first-page":"114","DOI":"10.3115\/1075812.1075835","article-title":"The Penn Treebank: Annotating predicate argument structure","volume-title":"Human Language Technology: Proceedings of a Workshop","author":"Marcus","year":"1994"},{"key":"2024092014245895000_bib26","first-page":"92","article-title":"Universal dependency annotation for multilingual parsing","volume-title":"Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)","author":"McDonald","year":"2013"},{"issue":"1959","key":"2024092014245895000_bib27","first-page":"1","article-title":"Dependency grammar and dependency parsing","volume":"5133","author":"Nivre","year":"2005","journal-title":"MSI report"},{"key":"2024092014245895000_bib28","article-title":"Introducing chatGPT","author":"OpenAI","year":"2023"},{"key":"2024092014245895000_bib29","first-page":"27730","article-title":"Training language models to follow instructions with human feedback","volume-title":"Advances in Neural Information Processing Systems","author":"Ouyang","year":"2022"},{"key":"2024092014245895000_bib30","first-page":"1","article-title":"Instruction tuning with GPT-4","author":"Peng","year":"2023","journal-title":"arXiv preprint arXiv:2304.03277"},{"key":"2024092014245895000_bib31","first-page":"257","article-title":"Multi-view Chinese treebanking","volume-title":"Proceedings of COLING 2014, the 25th International Conference on Computational Linguistics: Technical Papers","author":"Qiu","year":"2014"},{"key":"2024092014245895000_bib32","doi-asserted-by":"publisher","first-page":"695","DOI":"10.1162\/tacl_a_00294","article-title":"Deep contextualized self-training for low resource dependency parsing","volume":"7","author":"Rotman","year":"2019","journal-title":"Transactions of the Association for Computational Linguistics"},{"key":"2024092014245895000_bib33","doi-asserted-by":"publisher","first-page":"5004","DOI":"10.18653\/v1\/D18-1545","article-title":"Data augmentation via dependency tree morphing for low resource languages","volume-title":"2018 Conference on Empirical Methods in Natural Language Processing","author":"Sahin","year":"2018"},{"issue":"3","key":"2024092014245895000_bib34","doi-asserted-by":"publisher","first-page":"363","DOI":"10.1109\/TIT.1965.1053799","article-title":"Probability of error of some adaptive pattern-recognition machines","volume":"11","author":"Scudder","year":"1965","journal-title":"IEEE Transactions on Information Theory"},{"key":"2024092014245895000_bib35","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1186\/s40537-021-00492-0","article-title":"Text data augmentation for deep learning","volume":"8","author":"Shorten","year":"2021","journal-title":"Journal of Big Data"},{"issue":"398","key":"2024092014245895000_bib36","doi-asserted-by":"publisher","first-page":"528","DOI":"10.1080\/01621459.1987.10478458","article-title":"The calculation of posterior distributions by data augmentation","volume":"82","author":"Tanner","year":"1987","journal-title":"Journal of the American Statistical Association"},{"key":"2024092014245895000_bib37","article-title":"Stanford Alpaca: An instruction-following LLaMA model","author":"Taori","year":"2023"},{"key":"2024092014245895000_bib38","first-page":"1","article-title":"LLaMA: Open and efficient foundation language models","author":"Touvron","year":"2023","journal-title":"arXiv preprint arXiv:2302.13971"},{"key":"2024092014245895000_bib39","first-page":"1","article-title":"LLAMA 2: Open foundation and fine-tuned chat models","author":"Touvron","year":"2023","journal-title":"arXiv preprint arXiv:2307.09288"},{"issue":"1","key":"2024092014245895000_bib40","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1198\/10618600152418584","article-title":"The art of data augmentation","volume":"10","author":"Van Dyk","year":"2001","journal-title":"Journal of Computational and Graphical Statistics"},{"key":"2024092014245895000_bib41","author":"Vicuna","year":"2023"},{"key":"2024092014245895000_bib42","doi-asserted-by":"publisher","first-page":"5085","DOI":"10.18653\/v1\/2022.emnlp-main.340","article-title":"Super-Natural Instructions: Generalization via declarative instructions on 1600+ NLP tasks","volume-title":"Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing","author":"Wang","year":"2022"},{"key":"2024092014245895000_bib43","first-page":"24824","article-title":"Chain of thought prompting elicits reasoning in large language models","volume-title":"Advances in Neural Information Processing Systems","author":"Wei","year":"2022"},{"key":"2024092014245895000_bib44","doi-asserted-by":"publisher","first-page":"1","DOI":"10.18653\/v1\/2023.emnlp-main.44","article-title":"LLM-powered data augmentation for enhanced crosslingual performance","author":"Whitehouse","year":"2023","journal-title":"arXiv preprint arXiv:2305.14288"},{"issue":"2","key":"2024092014245895000_bib45","doi-asserted-by":"publisher","first-page":"207","DOI":"10.1017\/S135132490400364X","article-title":"The Penn Chinese Treebank: Phrase structure annotation of a large corpus","volume":"11","author":"Xue","year":"2005","journal-title":"Natural Language Engineering"},{"key":"2024092014245895000_bib46","doi-asserted-by":"publisher","first-page":"1","DOI":"10.18653\/v1\/W15-2201","article-title":"Domain adaptation for dependency parsing via self-training","volume-title":"Proceedings of the 14th International Conference on Parsing Technologies","author":"Yu","year":"2015"},{"key":"2024092014245895000_bib47","article-title":"mixup: Beyond empirical risk minimization","volume-title":"International Conference on Learning Representations","author":"Zhang","year":"2018"}],"container-title":["Computational Linguistics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/direct.mit.edu\/coli\/article-pdf\/50\/3\/867\/2470901\/coli_a_00515.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/direct.mit.edu\/coli\/article-pdf\/50\/3\/867\/2470901\/coli_a_00515.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,9,20]],"date-time":"2024-09-20T14:25:15Z","timestamp":1726842315000},"score":1,"resource":{"primary":{"URL":"https:\/\/direct.mit.edu\/coli\/article\/50\/3\/867\/120014\/LLM-Assisted-Data-Augmentation-for-Chinese"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024]]},"references-count":47,"journal-issue":{"issue":"3","published-online":{"date-parts":[[2024,9,1]]},"published-print":{"date-parts":[[2024,9,1]]}},"URL":"https:\/\/doi.org\/10.1162\/coli_a_00515","relation":{},"ISSN":["0891-2017","1530-9312"],"issn-type":[{"value":"0891-2017","type":"print"},{"value":"1530-9312","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2024]]},"published":{"date-parts":[[2024]]}}}