{"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/not-to-cry-wolf-distantly-supervised","title":"Not to Cry Wolf: Distantly Supervised Multitask Learning in Critical Care","arxiv_id":"1802.05027","date":"2018-02-14","proceeding":"ICML 2018 7","authors":["Patrick Schwab","Emanuela Keller","Carl Muroi","David J. Mack","Christian Strässle","Walter Karlen"],"abstract":"Patients in the intensive care unit (ICU) require constant and close\nsupervision. To assist clinical staff in this task, hospitals use monitoring\nsystems that trigger audiovisual alarms if their algorithms indicate that a\npatient's condition may be worsening. However, current monitoring systems are\nextremely sensitive to movement artefacts and technical errors. As a result,\nthey typically trigger hundreds to thousands of false alarms per patient per\nday - drowning the important alarms in noise and adding to the exhaustion of\nclinical staff. In this setting, data is abundantly available, but obtaining\ntrustworthy annotations by experts is laborious and expensive. We frame the\nproblem of false alarm reduction from multivariate time series as a\nmachine-learning task and address it with a novel multitask network\narchitecture that utilises distant supervision through multiple related\nauxiliary tasks in order to reduce the number of expensive labels required for\ntraining. We show that our approach leads to significant improvements over\nseveral state-of-the-art baselines on real-world ICU data and provide new\ninsights on the importance of task selection and architectural choices in\ndistantly supervised multitask learning.","url_abs":"http://arxiv.org/abs/1802.05027v2","url_pdf":"http://arxiv.org/pdf/1802.05027v2.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":"not-to-cry-wolf-distantly-supervised","repo_url":"https://github.com/d909b/DSMT-Nets","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[{"task_slug":"time-series-1","task_name":"Time Series"},{"task_slug":"time-series","task_name":"Time Series Analysis"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}