Am exploring what really can be said for AI in medicine. There are lots of good things going on … but some reality seems to have set in.
I ran into this conclusion in the paper Deep Learning for Genomics: A Concise Overview by Yue and Wang at Carnegie Mellon. [Yue, Tianwei and Haohan Wang. “Deep Learning for Genomics: A Concise Overview.” CoRR abs/1802.00810 (2018):]
Current applications, however, have not brought about a watershed revolution in genomic research. The predictive performances in most problems have not reach the expec- tation for real-world applications, neither have the interpretations of these abstruse models elucidate insightful knowledge. A plethora of new deep learning methods is constantly being proposed but awaits artful applications in genomics.
I was really hoping we were farther along. Maybe there’s hope … there’s always hope [Elvis: Farther along we’ll know more about it. Farther along we’ll understand why. Cheer up my brother live in the sunshine]. Right now, what I am seeing with Watson for Genomics, and other ‘production systems ‘ suggest lots of work ahead.