{"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/a-variational-approach-to-weakly-supervised","title":"A Variational Approach to Weakly Supervised Document-Level Multi-Aspect Sentiment Classification","arxiv_id":"1904.05055","date":"2019-04-10","proceeding":"NAACL 2019 6","authors":["Ziqian Zeng","Wenxuan Zhou","Xin Liu","Yangqiu Song"],"abstract":"In this paper, we propose a variational approach to weakly supervised\ndocument-level multi-aspect sentiment classification. Instead of using\nuser-generated ratings or annotations provided by domain experts, we use\ntarget-opinion word pairs as \"supervision.\" These word pairs can be extracted\nby using dependency parsers and simple rules. Our objective is to predict an\nopinion word given a target word while our ultimate goal is to learn a\nsentiment polarity classifier to predict the sentiment polarity of each aspect\ngiven a document. By introducing a latent variable, i.e., the sentiment\npolarity, to the objective function, we can inject the sentiment polarity\nclassifier to the objective via the variational lower bound. We can learn a\nsentiment polarity classifier by optimizing the lower bound. We show that our\nmethod can outperform weakly supervised baselines on TripAdvisor and\nBeerAdvocate datasets and can be comparable to the state-of-the-art supervised\nmethod with hundreds of labels per aspect.","url_abs":"http://arxiv.org/abs/1904.05055v1","url_pdf":"http://arxiv.org/pdf/1904.05055v1.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":"a-variational-approach-to-weakly-supervised","repo_url":"https://github.com/HKUST-KnowComp/VWS-DMSC","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"tf","reach":null}],"tasks":[{"task_slug":"classification","task_name":"General Classification"},{"task_slug":"sentiment-analysis","task_name":"Sentiment Analysis"},{"task_slug":"sentiment-classification","task_name":"Sentiment Classification"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1904.05055","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}