King's College London

Research portal

A neural candidate-selector architecture for automatic structured clinical text annotation

Research output: Chapter in Book/Report/Conference proceedingOther chapter contribution

Standard

A neural candidate-selector architecture for automatic structured clinical text annotation. / Singh, Gaurav; Marshall, Iain J.; Thomas, James; Shawe-Taylor, John; Wallace, Byron C.

CIKM 2017 - Proceedings of the 2017 ACM Conference on Information and Knowledge Management. Vol. Part F131841 Association for Computing Machinery, 2017. p. 1519-1528.

Research output: Chapter in Book/Report/Conference proceedingOther chapter contribution

Harvard

Singh, G, Marshall, IJ, Thomas, J, Shawe-Taylor, J & Wallace, BC 2017, A neural candidate-selector architecture for automatic structured clinical text annotation. in CIKM 2017 - Proceedings of the 2017 ACM Conference on Information and Knowledge Management. vol. Part F131841, Association for Computing Machinery, pp. 1519-1528, 26th ACM International Conference on Information and Knowledge Management, CIKM 2017, Singapore, Singapore, 6/11/2017. https://doi.org/10.1145/3132847.3132989

APA

Singh, G., Marshall, I. J., Thomas, J., Shawe-Taylor, J., & Wallace, B. C. (2017). A neural candidate-selector architecture for automatic structured clinical text annotation. In CIKM 2017 - Proceedings of the 2017 ACM Conference on Information and Knowledge Management (Vol. Part F131841, pp. 1519-1528). Association for Computing Machinery. https://doi.org/10.1145/3132847.3132989

Vancouver

Singh G, Marshall IJ, Thomas J, Shawe-Taylor J, Wallace BC. A neural candidate-selector architecture for automatic structured clinical text annotation. In CIKM 2017 - Proceedings of the 2017 ACM Conference on Information and Knowledge Management. Vol. Part F131841. Association for Computing Machinery. 2017. p. 1519-1528 https://doi.org/10.1145/3132847.3132989

Author

Singh, Gaurav ; Marshall, Iain J. ; Thomas, James ; Shawe-Taylor, John ; Wallace, Byron C. / A neural candidate-selector architecture for automatic structured clinical text annotation. CIKM 2017 - Proceedings of the 2017 ACM Conference on Information and Knowledge Management. Vol. Part F131841 Association for Computing Machinery, 2017. pp. 1519-1528

Bibtex Download

@inbook{3967c7967a6641b69be23a60dd7a4c51,
title = "A neural candidate-selector architecture for automatic structured clinical text annotation",
abstract = "We consider the task of automatically annotating free texts describing clinical trials with concepts from a controlled, structured medical vocabulary. Specifically, we aim to build a model to infer distinct sets of (ontological) concepts describing complementary clinically salient aspects of the underlying trials: the populations enrolled, the interventions administered and the outcomes measured, i.e., the PICO elements. This important practical problem poses a few key challenges. One issue is that the output space is vast, because the vocabulary comprises many unique concepts. Compounding this problem, annotated data in this domain is expensive to collect and hence sparse. Furthermore, the outputs (sets of concepts for each PICO element) are correlated: specific populations (e.g., diabetics) will render certain intervention concepts likely (insulin therapy) while effectively precluding others (radiation therapy). Such correlations should be exploited. We propose a novel neural model that addresses these challenges. We introduce a Candidate-Selector architecture in which the model considers setes of candidate concepts for PICO elements, and assesses their plausibility conditioned on the input text to be annotated. This relies on a 'candidate set' generator, which may be learned or relies on heuristics. A conditional discriminative neural model then jointly selects candidate concepts, given the input text. We compare the predictive performance of our approach to strong baselines, and show that it outperforms them. Finally, we perform a qualitative evaluation of the generated annotations by asking domain experts to assess their quality.",
author = "Gaurav Singh and Marshall, {Iain J.} and James Thomas and John Shawe-Taylor and Wallace, {Byron C.}",
year = "2017",
month = "11",
doi = "10.1145/3132847.3132989",
language = "English",
volume = "Part F131841",
pages = "1519--1528",
booktitle = "CIKM 2017 - Proceedings of the 2017 ACM Conference on Information and Knowledge Management",
publisher = "Association for Computing Machinery",

}

RIS (suitable for import to EndNote) Download

TY - CHAP

T1 - A neural candidate-selector architecture for automatic structured clinical text annotation

AU - Singh, Gaurav

AU - Marshall, Iain J.

AU - Thomas, James

AU - Shawe-Taylor, John

AU - Wallace, Byron C.

PY - 2017/11

Y1 - 2017/11

N2 - We consider the task of automatically annotating free texts describing clinical trials with concepts from a controlled, structured medical vocabulary. Specifically, we aim to build a model to infer distinct sets of (ontological) concepts describing complementary clinically salient aspects of the underlying trials: the populations enrolled, the interventions administered and the outcomes measured, i.e., the PICO elements. This important practical problem poses a few key challenges. One issue is that the output space is vast, because the vocabulary comprises many unique concepts. Compounding this problem, annotated data in this domain is expensive to collect and hence sparse. Furthermore, the outputs (sets of concepts for each PICO element) are correlated: specific populations (e.g., diabetics) will render certain intervention concepts likely (insulin therapy) while effectively precluding others (radiation therapy). Such correlations should be exploited. We propose a novel neural model that addresses these challenges. We introduce a Candidate-Selector architecture in which the model considers setes of candidate concepts for PICO elements, and assesses their plausibility conditioned on the input text to be annotated. This relies on a 'candidate set' generator, which may be learned or relies on heuristics. A conditional discriminative neural model then jointly selects candidate concepts, given the input text. We compare the predictive performance of our approach to strong baselines, and show that it outperforms them. Finally, we perform a qualitative evaluation of the generated annotations by asking domain experts to assess their quality.

AB - We consider the task of automatically annotating free texts describing clinical trials with concepts from a controlled, structured medical vocabulary. Specifically, we aim to build a model to infer distinct sets of (ontological) concepts describing complementary clinically salient aspects of the underlying trials: the populations enrolled, the interventions administered and the outcomes measured, i.e., the PICO elements. This important practical problem poses a few key challenges. One issue is that the output space is vast, because the vocabulary comprises many unique concepts. Compounding this problem, annotated data in this domain is expensive to collect and hence sparse. Furthermore, the outputs (sets of concepts for each PICO element) are correlated: specific populations (e.g., diabetics) will render certain intervention concepts likely (insulin therapy) while effectively precluding others (radiation therapy). Such correlations should be exploited. We propose a novel neural model that addresses these challenges. We introduce a Candidate-Selector architecture in which the model considers setes of candidate concepts for PICO elements, and assesses their plausibility conditioned on the input text to be annotated. This relies on a 'candidate set' generator, which may be learned or relies on heuristics. A conditional discriminative neural model then jointly selects candidate concepts, given the input text. We compare the predictive performance of our approach to strong baselines, and show that it outperforms them. Finally, we perform a qualitative evaluation of the generated annotations by asking domain experts to assess their quality.

UR - http://www.scopus.com/inward/record.url?scp=85037357460&partnerID=8YFLogxK

U2 - 10.1145/3132847.3132989

DO - 10.1145/3132847.3132989

M3 - Other chapter contribution

AN - SCOPUS:85037357460

VL - Part F131841

SP - 1519

EP - 1528

BT - CIKM 2017 - Proceedings of the 2017 ACM Conference on Information and Knowledge Management

PB - Association for Computing Machinery

ER -

View graph of relations

© 2018 King's College London | Strand | London WC2R 2LS | England | United Kingdom | Tel +44 (0)20 7836 5454