Description:

Xinlei Mi, PhD
Postdoctoral fellow
Department of Biostatistics
Columbia University 

Deep Learning in Biomedical Data

Big data in healthcare, including electronic medical records, large-scale clinical trials, high-dimensional omics data and etc, has grown rapidly and provided rich information to address various problems in bio-related areas. Machine learning and other modern statistical learning methods have been widely adopted in analyzing such data. We present two deep learning-based methods: deepTL for adjusting complex confounding, and Endlot in optimal individualized treatment rules. Comprehensive simulation studies and real applications are conducted to evaluate the proposed methods, both of which have shown efficient yet robust performance.