Charles Isbell makes the case for more ethical AI
The Robot Brains Podcast
00:00
The Unbalanced Datu Sati Jus of Labels
Machine learning is sprouting a whole new domain of research and how to optimize for that. We should be evaluating not just accuracy or class agnostic accuracy, but also things like truly out of out of sample data. You can't require that people be effectively perfect before you are willing to give them the ok to be one of the 20% who are going to be accepted for publication.
Transcript
Play full episode
