Non-Numerical Artificial Intelligence Implementation Learnings: Controlling Sampling Bias10/04/2017 Objective: Guidance As we look towards infusing machine learning into all parts of our lives, careful consideration must be given to the increasingly common problem of biased Artificial Intelligence. How can we ensure accuracy and mitigate subjectivity? Traditional methods to address bias have failed in our modern AI world. This talk will offer a new approach which uses the algorithms themselves to combat selection bias. Speaker(s)
|