520 karma · joined August 29, 2016
Produce follows a distribution along all characteristics, and that distribution changes over time. There's some you could discard a priori, but then there's the damage from customers manhandling it, the normal aging process, etc.
You can always decline to use a variable after the fact; she decided she didn't want to know the effect in the first place. Suppose the effect had been the opposite of what she suspected?
You can tell she's interested in preventing knowledge from how she handled her job to predict effectiveness of homelessness services - she actively decided to not use particular variables on the grounds they might show a result she was uncomfortable with. That isn't an issue of "being aware of the limitations of machine learning", it's intentional ignorance.
http://www.slate.com/articles/technology/future_tense/2016/0...
https://en.wikipedia.org/wiki/Chevron_U.S.A.,_Inc._v._Natura....
The ATF is particularly notorious for this.
OTOH there is generally enough flexibility for a court to punt when they don't actually want to address the underlying issue.
This is a reasonable law review article: http://www.tjeffersonlrev.org/sites/tjeffersonlrev.org/files...
That potentially includes the fully resolved types of objects (ie devirtualization), branch prediction (stronger than the CPU can do; for instance, if a value is only used inside a branch that's never taken, don't bother mutating it), data sizes (this "array" is only ever size 2, store it in registers), dead code elimination (keeps the compiled code small), and a whole bunch more fun stuff.
The historical solution (eg in the context of manorialism) was that if you couldn't afford to have a family, you didn't (and if you did, the kid got shunted into a monastery or the equivalent and was prevented from reproducing), so the "surplus" peasant population effectively shrank, and the English lower classes ended up being the descendants of downwardly-mobile gentry, who mostly found employment in different sectors.
That solution breaks down in the context of modern welfare systems.
The point of a UBI in the academic literature is to enable dismantling of high-overhead welfare state in favor of a low-overhead and higher-level grant (check out eg the overhead of Social Security Retirement income vs Social Security Disability income), but that tradeoff is somehow never actually offered.