-
Boosting the interpretability of clinical risk scores with intervention predictions
Authors:
Eric Loreaux,
Ke Yu,
Jonas Kemp,
Martin Seneviratne,
Christina Chen,
Subhrajit Roy,
Ivan Protsyuk,
Natalie Harris,
Alexander D'Amour,
Steve Yadlowsky,
Ming-Jun Chen
Abstract:
Machine learning systems show significant promise for forecasting patient adverse events via risk scores. However, these risk scores implicitly encode assumptions about future interventions that the patient is likely to receive, based on the intervention policy present in the training data. Without this important context, predictions from such systems are less interpretable for clinicians. We prop…
▽ More
Machine learning systems show significant promise for forecasting patient adverse events via risk scores. However, these risk scores implicitly encode assumptions about future interventions that the patient is likely to receive, based on the intervention policy present in the training data. Without this important context, predictions from such systems are less interpretable for clinicians. We propose a joint model of intervention policy and adverse event risk as a means to explicitly communicate the model's assumptions about future interventions. We develop such an intervention policy model on MIMIC-III, a real world de-identified ICU dataset, and discuss some use cases that highlight the utility of this approach. We show how combining typical risk scores, such as the likelihood of mortality, with future intervention probability scores leads to more interpretable clinical predictions.
△ Less
Submitted 6 July, 2022;
originally announced July 2022.
-
Concept-based model explanations for Electronic Health Records
Authors:
Diana Mincu,
Eric Loreaux,
Shaobo Hou,
Sebastien Baur,
Ivan Protsyuk,
Martin G Seneviratne,
Anne Mottram,
Nenad Tomasev,
Alan Karthikesanlingam,
Jessica Schrouff
Abstract:
Recurrent Neural Networks (RNNs) are often used for sequential modeling of adverse outcomes in electronic health records (EHRs) due to their ability to encode past clinical states. These deep, recurrent architectures have displayed increased performance compared to other modeling approaches in a number of tasks, fueling the interest in deploying deep models in clinical settings. One of the key ele…
▽ More
Recurrent Neural Networks (RNNs) are often used for sequential modeling of adverse outcomes in electronic health records (EHRs) due to their ability to encode past clinical states. These deep, recurrent architectures have displayed increased performance compared to other modeling approaches in a number of tasks, fueling the interest in deploying deep models in clinical settings. One of the key elements in ensuring safe model deployment and building user trust is model explainability. Testing with Concept Activation Vectors (TCAV) has recently been introduced as a way of providing human-understandable explanations by comparing high-level concepts to the network's gradients. While the technique has shown promising results in real-world imaging applications, it has not been applied to structured temporal inputs. To enable an application of TCAV to sequential predictions in the EHR, we propose an extension of the method to time series data. We evaluate the proposed approach on an open EHR benchmark from the intensive care unit, as well as synthetic data where we are able to better isolate individual effects.
△ Less
Submitted 8 March, 2021; v1 submitted 3 December, 2020;
originally announced December 2020.
-
Catalogue of star positions and B-magnitudes in 60th declination zone based on UkrVO Joint Digital Archive
Authors:
V. N. Andruk,
L. K. Pakuliak,
V. V. Golovnya,
G. O. Ivanov,
O. M. Yizhakevych,
Yu. I. Protsyuk,
S. V. Shatokhina
Abstract:
Catalogue of star positions and B-magnitudes based on UkrVO Joint Digital Archive has been created in 60-degree declination zone of FONAK(FON) observational program as the first attempt to use the commercial scanners for astrometri c purposes. The height of zone is 8 degree, number of involved plates is 120. Digital images of plates were obtained using Microtek ScanMaker 9800XL TMA commercial scan…
▽ More
Catalogue of star positions and B-magnitudes based on UkrVO Joint Digital Archive has been created in 60-degree declination zone of FONAK(FON) observational program as the first attempt to use the commercial scanners for astrometri c purposes. The height of zone is 8 degree, number of involved plates is 120. Digital images of plates were obtained using Microtek ScanMaker 9800XL TMA commercial scanner, with plate resolution 1200 dpi, linear dimensions 13,000x13,000 px for plates 30x30cm. The catalogue includes 1, 263, 932 stars and galaxies down to B 16.5 mag at the epoch 1984.76+/-0.50. Positions of objects are in TYCHO-2 reference frame, B-magnitudes in the system of photoelectric standards. The internal accuracy for all objects on both coordinates RA,DEC is +/-0.26 arcsec.The accuracy of B magnitudes determination is +/-0.17 mag, except stars in the B-interval 8-13 mag having the positional errors +/-0.13 arcsec and photometric ones +/-0.11 mag. The convergence of star positions with TYCHO-2 system is +/-0.06 arcsec on both coordinates(based on 93,925 stars), the convergence with photoelectric standards system is +/-0.16 mag (based on 4,458 stars). External comparison with UCAC-4 gives the positional errors +/-0.34 arcsec based on 1, 099, 005 cross-identified objects.
△ Less
Submitted 17 December, 2015;
originally announced December 2015.
