Math against crimes against humanity: Using rigorous statistics to prove genocide when the dead cannot speak for themselves

[doctorow]

Originally published at: https://boingboing.net/2019/01/01/counting-the-dead.html

…

[Boundegar]

There’s no witness you can kill to silence a machine-learning model.

Yes there is. If the model is developed by your own citizen, you can put him right in the ground. If they’re not a citizen, they’re “outside agitators.” Use ad-hominem attacks to discredit the research, hire some academics to express doubt, and rely on the press to ignore anything involving numbers.

[stephen_schenck]

at least a 65 percent chance of harboring graves

That … doesn’t sound particularly compelling?

[Ratel]

Awesome. Let’s start with Iraq.

[Bernel]

@Ratel

Awesome. Let’s start with Iraq.

You don’t need graves when you have big enough bombs.

[roomwithaview]

@stephen_schenck

at least a 65 percent chance of harboring graves

That … doesn’t sound particularly compelling?

*It’s infinitely greater than 0%


*_{See also: why the press doesn’t like numbers}

[Nova]

random forests is not exactly rigorous statistics, but I still think it’s a worthy approach.