Accepted author manuscript, 517 KB, PDF document
Final published version
Research output: Contribution in Book/Report/Proceedings - With ISBN/ISSN › Conference contribution/Paper › peer-review
Research output: Contribution in Book/Report/Proceedings - With ISBN/ISSN › Conference contribution/Paper › peer-review
}
TY - GEN
T1 - AdaCare
T2 - Explainable Clinical Health Status Representation Learning via Scale Adaptive Feature Extraction and Recalibration
AU - Ma, Liantao
AU - Gao, Junyi
AU - Wang, Yasha
AU - Wang, Jiangtao
AU - Ruan, Wenjie
AU - Zhang, Chaohe
AU - Tang, Wen
AU - Gao, Xin
AU - Ma, Xinyu
PY - 2020/4/3
Y1 - 2020/4/3
N2 - Deep learning-based health status representation learning and clinical prediction have raised much research interest in recent years. Existing models have shown superior performance, but there are still several major issues that have not been fully taken into consideration. First, the historical variation pattern of the biomarker in diverse time scales plays an important role in indicating the health status, but it has not been explicitly extracted by existing works. Second, key factors that strongly indicate the health risk are different among patients. It is still challenging to adaptively make use of the features for patients in diverse conditions. Third, using the prediction model as a black box will limit the reliability in clinical practice. However, none of the existing works can provide satisfying interpretability and meanwhile achieve high prediction performance. In this work, we develop a general health status representation learning model, named AdaCare. It can capture the long and short-term variations of biomarkers as clinical features to depict the health status in multiple time scales. It also models the correlation between clinical features to enhance the ones which strongly indicate the health status and thus can maintain a state-of-the-art performance in terms of prediction accuracy while providing qualitative in- interpretability. We conduct health risk prediction experiment on two real-world datasets. Experiment results indicate that AdaCare outperforms state-of-the-art approaches and provides effective interpretability which is verifiable by clinical experts.
AB - Deep learning-based health status representation learning and clinical prediction have raised much research interest in recent years. Existing models have shown superior performance, but there are still several major issues that have not been fully taken into consideration. First, the historical variation pattern of the biomarker in diverse time scales plays an important role in indicating the health status, but it has not been explicitly extracted by existing works. Second, key factors that strongly indicate the health risk are different among patients. It is still challenging to adaptively make use of the features for patients in diverse conditions. Third, using the prediction model as a black box will limit the reliability in clinical practice. However, none of the existing works can provide satisfying interpretability and meanwhile achieve high prediction performance. In this work, we develop a general health status representation learning model, named AdaCare. It can capture the long and short-term variations of biomarkers as clinical features to depict the health status in multiple time scales. It also models the correlation between clinical features to enhance the ones which strongly indicate the health status and thus can maintain a state-of-the-art performance in terms of prediction accuracy while providing qualitative in- interpretability. We conduct health risk prediction experiment on two real-world datasets. Experiment results indicate that AdaCare outperforms state-of-the-art approaches and provides effective interpretability which is verifiable by clinical experts.
U2 - 10.1609/aaai.v34i01.5427
DO - 10.1609/aaai.v34i01.5427
M3 - Conference contribution/Paper
SN - 9781577358350
T3 - Proceedings of the AAAI Conference on Artificial Intelligence
SP - 825
EP - 832
BT - The Thirty-Fourth AAAI Conference on Artificial Intelligence, 2020
PB - AAAI
ER -