gbrown

57 ms·

511.
by gbrown 7y ago
> The standard best practices in STATS 101 is to compute R^2 coefficient Any regression class which doesn't teach about overfitting is incomplete. R^2 is still a very useful measure, however, if you've come up with a parsimonio
512.
by gbrown 7y ago
No, GPUs are quite different than CPUs with more cores. They're optimized for Singl-Instruction-Multiple-Data algorithms, in which the same operation is done at the same time to many different inputs (think image or video processing).
513.
by gbrown 8y ago
Then why are wealthy people paying (and in many cases risking) so much to get their kids into elite schools? It's an important signalling complement to existing networks, and a way to extend them.
514.
by gbrown 8y ago
Elite schools are a great mechanism to entrench privelege. Paying to get your slacker kid a middling GPA at an elite school is a great way to connect them to a business network of other wealthy people, and fast track them for high paying jo
515.
by gbrown 8y ago
"signatures", but how good were they? I'd not be surprised if you could do much better than random guesswork, but I'd need to see a lot more data to accept that this is good enough to be scaled up population-wide.
516.
by gbrown 8y ago
Those are the examples given, but that wouldn't explain the huge numbers claimed elsewhere - nearly a quarter of phones in the US. Pardon me for not assuming that a data gathering corporation has my best interests at heart.
517.
by gbrown 8y ago
Perhaps we could imagine better privacy laws. The fact that something is legal, and a legitimate business, doesn't mean that it's good or should be allowed.
518.
by gbrown 8y ago
First, it's not clear to me that this is what is being proposed - instead, it sounds like the companies involved get their surveillance tech rolled into other apps like adware. Second, even if enrollment in the program was done more ca
519.
by gbrown 8y ago
Don't get me wrong, I bike too. I also have little faith that this sort of system will be implemented effectively and transparently. I also don't think it should be legal in the first place to gather this kind of data without info
520.
by gbrown 8y ago
Are they accounting for carpooling/passengers? I drive into work with my wife every morning, and often check my phone en-route while she's driving. This is an important problem, but ham fisted solutions which invade privacy are no
521.
by gbrown 8y ago
> I’m saying if I’m looking at the data My point is about the data generating mechanism - making inferences conditional on a bad selection procedure is a bad idea. It would be bad statistics, and it's bad ML. In this case, my argume
522.
by gbrown 8y ago
I think your attitude is really, really dangerous, because it ignores the long term effects these kinds of things have on society. People don't even really understand how humans apply rules in biased ways - when you attache the words
523.
by gbrown 8y ago
So "legal" is the threshold for "bad"? I disagree. If the proportion of white shoplifters and black shoplifters being caught is different, AI will just scale up that disparity. If reality is biased, we should work to fi
524.
by gbrown 8y ago
I mean, that's why people do theoretical physics in higher dimensions and then search desperately for ways to test their theories. We can extrapolate with mathematics in infinite ways, but until we can find ways to test things it'
525.
by gbrown 8y ago
I started typing a reply, but as usual Wikipedia has a better response than I do: https://en.wikipedia.org/wiki/Occam%27s_razor#Justifications
526.
by gbrown 8y ago
Not a physicist, but at a guess: Occam's Razor, plus the existance of simpler theories which explain our current observations.
527.
by gbrown 8y ago
Exactly
528.
by gbrown 8y ago
We can see all that from here though.
529.
by gbrown 8y ago
If you're thinking of the LMC as an "intersting place", you're committing a pretty huge error of scale.
530.
by gbrown 8y ago
In high school we messed around with this in the crappy P.O.S. system at the place I worked (windowed app, running on Windows) to see what data was recorded on all our various student ID cards, gift certificates etc.
531.
by gbrown 8y ago
I hate it when colleagues and students hand me a USB stick to use. We have great file sharing infrastructure, there's no reason for me to plug in your USB stick to access some powerpoint you want me to look at. Now get off my lawn.
532.
by gbrown 8y ago
It does exist in English.
533.
by gbrown 8y ago
It's still really convoluted and removed from the actual elephant. From a frequentist perspective, the elephant either is or is not on the table, and if it is then it has a certain size - not random quantities subject to probabilistic
534.
by gbrown 8y ago
My first thought as well. Typed from my s̶u̶r̶v̶e̶i̶l̶l̶a̶n̶c̶e̶ ̶d̶e̶v̶i̶c̶e̶ smartphone.
535.
by gbrown 8y ago
I can second the use of audacity for sound editing in Kdenlive. I found it pretty stable though (2018).
536.
by gbrown 8y ago
The authors note that judges may care explicitly about racial bias, but based on a quick read they're making a really, really big mistake in the language they're using: they confuse arrests with crime. Arrests and convictions are
537.
by gbrown 8y ago
If I recall correctly, such a treatment is not yet evidence based, though clinical trials are ongoing.
538.
by gbrown 8y ago
If only - expect Florida heat, Canada cold, and extreme weather of all types.
539.
by gbrown 8y ago
Still, that kills any incentive for fee competition between cards - they all share the benefits of being convenient and face little pressure to change the fee structure.
540.
by gbrown 8y ago
There was just a lawsuit about that, and it went... poorly. https://www.washingtonpost.com/business/economy/supreme-cour...

More ›