Limitations of Deep Learning Attention Mechanisms in Clinical Research: Empirical Case Study Based on the Korean Diabetic Disease Setting.

Journal: Journal of medical Internet research
Published Date:

Abstract

BACKGROUND: Despite excellent prediction performance, noninterpretability has undermined the value of applying deep-learning algorithms in clinical practice. To overcome this limitation, attention mechanism has been introduced to clinical research as an explanatory modeling method. However, potential limitations of using this attractive method have not been clarified to clinical researchers. Furthermore, there has been a lack of introductory information explaining attention mechanisms to clinical researchers.

Authors

  • Junetae Kim
    Graduate School of Cancer Science and Policy, National Cancer Center, Goyang-si, Gyeonggi-do, Republic of Korea.
  • Sangwon Lee
    Urban Robotics Laboratory (URL), Dept. Civil and Environmental Engineering, Korea Advanced Institute of Science and Technology (KAIST), 291 Daehak-ro, Yuseong-gu, Daejeon 305-338, Korea. lsw618@gmail.com.
  • Eugene Hwang
    School of Management Engineering, Korea Advanced Institute of Science and Technology, Seoul, Republic of Korea.
  • Kwang Sun Ryu
    Cancer Data Center, National Cancer Control Institute, National Cancer Center, Goyang-si, Gyeonggi-do, Republic of Korea.
  • Hanseok Jeong
    Graduate School of Cancer Science and Policy, National Cancer Center, Goyang-si, Gyeonggi-do, Republic of Korea.
  • Jae Wook Lee
    Department of Radiology, Soonchunhyang University Bucheon Hospital, Gyeonggi-do, Korea.
  • Yul Hwangbo
    Healthcare AI Team, Healthcare Platform Center, National Cancer Center, Goyang-si, Gyeonggi-do, Republic of Korea.
  • Kui Son Choi
    Graduate School of Cancer Science and Policy, National Cancer Center, Goyang-si, Gyeonggi-do, Republic of Korea.
  • Hyo Soung Cha
    Graduate School of Cancer Science and Policy, National Cancer Center, Goyang-si, Gyeonggi-do, Republic of Korea.