You can add location information to your Tweets, such as your city or precise location, from the web and via third-party applications. You always have the option to delete your Tweet location history. Learn more
I’m confused. Algorithms fairness techniques like fair empirical risk minimization have proofs showing that the models they produce are fair by accepted mathematical definitions like equal opportunity. What’s wrong with these approaches?
Challenge is that these math definitions are created by #AI researchers not social scientists and policy makers, who are experts. Also, these constrained models are only shown on small scale. As you scale up models, accuracy severely suffers, because we don't have good techniques
Agreed, but I fear we're letting the perfect be the enemy of the good. These techniques may be limited to simple/linear models but that's a lot of real-world ML, where bias is hurting ppl today. I also suspect we won't find the best objective funcs w/o testing these in practice.
Most agree that "ML bias is a serious problem" but how do we act on that? Realistically few will slow down use of ML. I think we should encourage everybody to try 1st-order/approx techniques of algorithmic fairness to see in practice how they help, and learn how to improve them.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.
