{"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/strong-baselines-for-neural-semi-supervised","title":"Strong Baselines for Neural Semi-supervised Learning under Domain Shift","arxiv_id":"1804.09530","date":"2018-04-25","proceeding":"ACL 2018 7","authors":["Sebastian Ruder","Barbara Plank"],"abstract":"Novel neural models have been proposed in recent years for learning under\ndomain shift. Most models, however, only evaluate on a single task, on\nproprietary datasets, or compare to weak baselines, which makes comparison of\nmodels difficult. In this paper, we re-evaluate classic general-purpose\nbootstrapping approaches in the context of neural networks under domain shifts\nvs. recent neural approaches and propose a novel multi-task tri-training method\nthat reduces the time and space complexity of classic tri-training. Extensive\nexperiments on two benchmarks are negative: while our novel method establishes\na new state-of-the-art for sentiment analysis, it does not fare consistently\nthe best. More importantly, we arrive at the somewhat surprising conclusion\nthat classic tri-training, with some additions, outperforms the state of the\nart. We conclude that classic approaches constitute an important and strong\nbaseline.","url_abs":"http://arxiv.org/abs/1804.09530v1","url_pdf":"http://arxiv.org/pdf/1804.09530v1.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":"strong-baselines-for-neural-semi-supervised","repo_url":"https://github.com/bplank/semi-supervised-baselines","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"strong-baselines-for-neural-semi-supervised","repo_url":"https://github.com/ambujojha/SemiSupervisedLearning","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"domain-adaptation","task_name":"Domain Adaptation"},{"task_slug":"multi-task-learning","task_name":"Multi-Task Learning"},{"task_slug":"part-of-speech-tagging","task_name":"Part-Of-Speech Tagging"},{"task_slug":"sentiment-analysis","task_name":"Sentiment Analysis"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/sentiment-analysis-on-multi-domain-sentiment","task":"Sentiment Analysis","dataset":"Multi-Domain Sentiment Dataset","model":"Multi-task tri-training","rank_in_archive_order":3,"of":6,"metrics":{"Average":"79.15","Books":"74.86","DVD":"78.14","Electronics":"81.45","Kitchen":"82.14"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1804.09530","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}