迁移学习在医疗卫生领域中的应用

Application of Transfer Learning in Medical and Healt h Field

迁移学习是一种新兴的机器学习技术,如果能够合理地将迁移学习技术应用在医疗卫生领域中,则可以有效解决传统机器学习和深度学习方法在医疗领域中所需数据标签不足的缺陷。笔者介绍了迁移学习的基本概念、迁移学习与传统机器学习方法之间的区别,同时从医学影像、医学文本、医学信息资源以及安全问题4个方面,介绍了近几年国内研究人员提出的几种将迁移学习应用在医疗卫生领域的方法,对于相关领域的研究人员有一定的借鉴价值。

Transfer learning is an emerging machine learning technology.If the transfer learning technology can be reasonably Applied in the medical and health field,it can effectively solve the defect of insufficient data labels required by traditional machine learning and deep learning methods in the medical field.The author introduces the basic concepts of transfer learning and the difference between transfer learning and traditional machine learning methods.At the same time,starting from the four aspects of medical imaging,medical texts,medical information resources,and security issues,it introduces what domestic researchers have put forward in recent years.Several methods of Applying transfer learning in the medical and health field have certain reference value for researchers in related fields.