Recommended: Microsoft is creating an oracle for catching biased AI algorithms

Artificial Intelligence (AI) algorithms that are used for crime detection, loan approvals, and employee evaluations are considered by many to be objective, but they can sometimes have many of the same prejudices and biases that human evaluators have. Given the opacity of many black box approaches to AI, this could lead to serious problems with fairness and equity. This article discusses an admittedly imperfect approach by Microsoft to evaluate these AI algorithms using (surprise!) an AI algorithm. It flags situations where an algorithm appears to have problems with unfair differential treatments  based on race, gender, or age. Continue reading →

PMean: What are we doing to justify all that time we’re budgeting?

An email discussion about the appropriate percentage effort on research grants has produced a lot of interesting discussions. One person raised an interesting question. The typical data analysis, he claimed, might involve a few hours reviewing the input data set, a few hours conducting the analysis and a few hours preparing a statistical summary, but even after a generous estimate of the work at each of the time points, he could only come up with 22 hours of effort, which corresponds roughly with a 1% FTE. I wrote back describing some of the things that might occur before the data analysis that might add time to this effort. Continue reading →