Header

UZH-Logo

Maintenance Infos

Not to Cry Wolf: Distantly Supervised Multitask Learning in Critical Care


Schwab, Patrick; Keller, Emanuela; Muroi, Carl; Mack, David J; Strässle, Christian; Karlen, Walter (2018). Not to Cry Wolf: Distantly Supervised Multitask Learning in Critical Care. In: 35 th International Conference on Machine Learning, Stockholm, Sweden, 10 July 2018 - 15 July 2018.

Abstract

Patients in the intensive care unit (ICU) require constant and close supervision. To assist clinical staff in this task, hospitals use monitoring systems that trigger audiovisual alarms if their algorithms indicate that a patient's condition may be worsening. However, current monitoring systems are extremely sensitive to movement artefacts and technical errors. As a result, they typically trigger hundreds to thousands of false alarms per patient per day - drowning the important alarms in noise and adding to the exhaustion of clinical staff. In this setting, data is abundantly available, but obtaining trustworthy annotations by experts is laborious and expensive. We frame the problem of false alarm reduction from multivariate time series as a machine-learning task and address it with a novel multitask network architecture that utilises distant supervision through multiple related auxiliary tasks in order to reduce the number of expensive labels required for training. We show that our approach leads to significant improvements over several state-of-the-art baselines on real-world ICU data and provide new insights on the importance of task selection and architectural choices in distantly supervised multitask learning.

Abstract

Patients in the intensive care unit (ICU) require constant and close supervision. To assist clinical staff in this task, hospitals use monitoring systems that trigger audiovisual alarms if their algorithms indicate that a patient's condition may be worsening. However, current monitoring systems are extremely sensitive to movement artefacts and technical errors. As a result, they typically trigger hundreds to thousands of false alarms per patient per day - drowning the important alarms in noise and adding to the exhaustion of clinical staff. In this setting, data is abundantly available, but obtaining trustworthy annotations by experts is laborious and expensive. We frame the problem of false alarm reduction from multivariate time series as a machine-learning task and address it with a novel multitask network architecture that utilises distant supervision through multiple related auxiliary tasks in order to reduce the number of expensive labels required for training. We show that our approach leads to significant improvements over several state-of-the-art baselines on real-world ICU data and provide new insights on the importance of task selection and architectural choices in distantly supervised multitask learning.

Statistics

Citations

Downloads

16 downloads since deposited on 18 Feb 2019
14 downloads since 12 months
Detailed statistics

Additional indexing

Item Type:Conference or Workshop Item (Paper), not_refereed, original work
Communities & Collections:04 Faculty of Medicine > University Hospital Zurich > Clinic for Neurosurgery
Dewey Decimal Classification:610 Medicine & health
Language:English
Event End Date:15 July 2018
Deposited On:18 Feb 2019 15:24
Last Modified:30 Oct 2019 08:13
Publisher:International Conference on Machine Learning
OA Status:Green
Official URL:https://icml.cc/Conferences/2018/
Other Identification Number:arXiv:1802.05027

Download

Green Open Access

Download PDF  'Not to Cry Wolf: Distantly Supervised Multitask Learning in Critical Care'.
Preview
Content: Published Version
Filetype: PDF
Size: 1MB