currymj

49 ms·

601.
by currymj 7y ago
although it’s probably in appendix D due to page limits.
602.
by currymj 7y ago
AI has a ton of impossibility results. the "no free lunch theorems" are probably the best known. people now actively think in terms of designing assumptions, priors, or inductive biases to make their systems work better on the dat
603.
by currymj 7y ago
i have noticed that to get day-to-day mathematical work done of any kind, you basically have to at least pretend to be a mathematical Platonist. the mathematical objects are things "out there" to be discovered or manipulated. ever
604.
by currymj 7y ago
usually a good-sounding idea like this, if it hasn't been tried, must have some hidden pitfalls. but the number of people writing parser-based interactive fiction is ludicrously small, so probably just nobody thought to do it.
605.
by currymj 7y ago
you can definitely tell people not to do certain things or use certain technologies and it mostly works. as a simple example, sometimes when getting access to a dataset, you have to sign a contract agreeing not to try to de-anonymize or ide
606.
by currymj 7y ago
the YoloV3 (the model that these tools are designed to work with) paper is extremely funny and worth reading for anyone who hasn't. https://pjreddie.com/media/files/papers/YOLOv3.pdf
607.
by currymj 7y ago
they're hyper-selective, but there's no particular reason to arbitrarily fail 1/3 of their students. you could just make the classes harder until this happens, but employers/grad schools seem to be pretty happy the way t
608.
by currymj 7y ago
most of the standard metrics of fairness for machine learning don't just just try to equalize proportions of positive/negative labels. they look at error rates. under these measures of fairness, a perfectly accurate predictor is r
609.
by currymj 7y ago
I really doubt that Lyndon Johnson was thinking about how much he hates men when he signed the Fair Housing Act into law.
610.
by currymj 7y ago
this is a pretty paranoid view. the restrictions only apply to very specific categories where there have been laws against discrimination for many decades. perfume and razors are not among them.
611.
by currymj 7y ago
if you go look at the twitter accounts of the journalists you can see all the gossip in real time. it seems like a lot of people are very unhappy about the Bloomberg reporting decisions. (there are occasionally actual articles written about
612.
by currymj 7y ago
i think the attitude is that if you theoretically could be making more money, then you are actually losing money.
613.
by currymj 7y ago
this seems to be just a copy-and-pasted article from the New York Times. https://www.nytimes.com/2019/12/04/us/politics/cia-torture-d...
614.
by currymj 7y ago
typically the actual papers are reasonably honest about everything and the misinformation comes in elsewhere, often university press releases summarizing the results.
615.
by currymj 7y ago
Disturbingly common, according to a NYT investigation. Moreover the machines are often improperly calibrated, software is not available, etc. https://www.nytimes.com/2019/11/03/business/drunk-driving-br..
616.
by currymj 7y ago
as early as the 90s, automated control systems in F1 resulted in cars (i.e. Williams FW14) that to some extent drove themselves better than any human. indeed, many of the systems used have been specifically banned since then.
617.
by currymj 7y ago
i wouldn't underestimate the power of good tools here. All the software libraries for ML are very easy to get started with, and make it very easy to prototype cool things. It seems like in other applied areas it's a lot more work
618.
by currymj 7y ago
this is how most of the constrained MDP stuff effectively works, it’s not a bad intuition that it is just different kinds of reward shaping. in some approaches you write down the Lagrangian of the RL reward-maximizing problem and then the h
619.
by currymj 7y ago
when your brain gets to this point, just log off and go look at a tree or something. this is serious overexposure to twitter.
620.
by currymj 7y ago
ironically, Hawaii has no death penalty and extremely restrictive gun laws. it's a big country!
621.
by currymj 7y ago
a posteriori, though, it is in fact extremely common for people to suffer from these symptoms after having those experiences.
622.
by currymj 7y ago
of course everyone mostly knows this but for some reason the people making the rankings want a fig leaf of objectivity, so they come up with a weird formula. and then the formula can always be gamed.
623.
by currymj 7y ago
I wish university rankings didn't exist, or at least were never taken seriously by anyone. Failing that, I wish the people making the rankings would just outright admit they're making a subjective list, so at least universities wo
624.
by currymj 7y ago
I am a native English speaker in the US and to me it seems palpably unfair. People much smarter than me have to spend a lot of time struggling with a foreign language, and are still unfairly perceived as less intelligent than native speaker
625.
by currymj 7y ago
one thing I think about a lot is that Australia has the sort of voting system that voting system nerds here in the US dream about. yet their political situation seems to be about as dysfunctional as ours. of course first past the post is a
626.
by currymj 7y ago
this is equally true of, say, schizophrenia — there’s no physical test for it as far as I know, just questionnaires about how someone perceives the world. but i don’t think anyone would argue it’s not a biological disease at this point, eve
627.
by currymj 7y ago
use it every day, very pleasant distribution for a scientific workstation, especially the great nvidia driver support. the overall design seems to be for a knowledgeable computer user, but one who has no interest in configuring things.
628.
by currymj 7y ago
generally speaking it seems like it is insider trading in Europe and is not in the US.
629.
by currymj 7y ago
there is an organization called Software Carpentry that is basically like this. however, the real problem has more to do with incentives and funding than it does knowledge of software best practices. right now, if you take the time to write
630.
by currymj 7y ago
could this not just be a weird dude who happens to be funded by the Navy in some more mundane capacity, and therefore assigns patents to the Navy?

More ›