@moorejh 3 years ago
The labels attached to images used to train machine-vision systems are often wrong. That could mean bad decisions by self-driving cars and medical algorithms. https://www.wired.com/story/foundations-ai-riddled-errors/ #machinelearning #datascience #deeplearning
@ewenharrison 4 years ago
Black box models are not acceptable for high stakes decisions in medicine. Models must be interpretable, not just explainable (explanations are wrong by definition). Here’s your great long-read for today by @CynthiaRudin https://arxiv.org/abs/1811.10154 #rstats #DataScience