Home
> Uncategorized > Teaching Computers to be Fair
Teaching Computers to be Fair
March 15, 2018
My newest Bloomberg View piece just came out:
How to Teach a Computer What ‘Fair’ Means
If we’re going to rely on algorithms, we’ll have to figure it out.
To read all of my Bloomberg View pieces, go here.
Categories: Uncategorized
mathbabe 
What a progressive defines as fair, a conservative defines as unfair. Same in the other direction. Take something simple as affirmative action / diversity. We will never get agreement on what’s fair, because as soon as you advantage one group, you disadvantage another. So fair ultimately will be decided by whoever has political control at the moment.
LikeLike
“The challenge, then, is to get tech giants and others to recognize — and take responsibility for — the effects their algorithms can have on society.”
Good article and I am glad you are deciphering what algorithms have to say and don’t say.
President Lyndon Johnson knew a bit about deciphering.
“The most important thing a man has to tell you is what he’s not telling you,” he said. “The most important thing he has to say is what he’s trying not to say.”
― Robert A. Caro, Master of the Senate
LikeLike
The good news is that as the corporations gather more and more data, they will be able to better provide individual results instead of relying on the broader strokes that are still used. It’s also going to be harder and harder to determine what’s ‘fair’ or how to tweak it, Neural Networks routines are notoriously hard to debug, they aren’t a ‘simple’ formula like the FICO score. What happens when the Neural network that includes the web site history notices that you visit Spanish speaking sites and uses that as part of the credit score? It wasn’t coded that way, it just happened as the scoring grew more and more complex to be as accurate as possible.
LikeLike
i just saw this…
https://www.theguardian.com/news/2018/mar/17/data-war-whistleblower-christopher-wylie-faceook-nix-bannon-trump
LikeLike