/sqt/ Stupid Questions Thread: Barry Edition

This thread is for questions that don't deserve their own thread.

Tips!
>give context
>describe your thought process if you're stuck
>try wolframalpha.com and stackexchange.com
>How To Ask Questions The Smart Way catb.org/~esr/faqs/smart-questions.html

Previous thread

Other urls found in this thread:

projecteuler.net/problem=46
avxhm.se/ebooks/12920980661292098066.html
twitter.com/NSFWRedditVideo

Why am I so shit at life?

how. I assume I have to use expectation maximisation but how

You don't meditate.

>how. I assume I have to use expectation maximisation but how
What have you tried?

well for now I'm rereading the notes of expectation maximization, the part I don't understand is where the additional L2 norm would fit in in all this

Should I ask about neural networks and back propagation here?

>neural networks and back propagation
not science or math

recommend a brainlet (me) a book on theoretical physics. not for a specific subject, a text that gives me an idea of how theoretical works. something that explains why we integrate here or use a fourier transform there. that kind of stuff

But /g/ is a shit board. Where should I look instead?

nevermind I'm dumb, all I have to do for this exercise is derivate the given equation w.r.t u_k, it's actually easy as fuck

is STEM even worth anything anymore? I'm in EE and I'm probably gonna finish my program, but I'm so fucking cynical when it comes to employment prospects after graduation.

>he doesn't understand neural networks
Yes they most certainly are both science and math, so fuck off

What is your question user this is the correct board, /g/ is for technology and there are some programming kiddos but here is where the Computer Scientists lurk.

Even if butthurt engineers like
Say different

>>he doesn't understand neural networks
I'm not a "he".

fuck off faggot

>fuck off faggot
Why the homophobia?

>>he doesn't understand neural networks
I'm not a "neural networks".

>i'm not a "he"

honestly it was interesting the first 10 times, by now you're just a tiresome fucking faggot.

>honestly it was interesting the first 10 times, by now you're just a tiresome fucking faggot.
Your homophobia was never interesting.

okay buddy I'll still play with you
Why the faggotry?

>physics 1
>find the time of a ball thrown in the air that reaches a max height of 5000feet.
>spend 15minutes thinking about it
>cant figure it out
>look up answer
>set vf to 0 at its peak

if i didnt realize that, am i fucked the rest of the semester?

>okay buddy I'll still play with you
>Why the faggotry?
Are you okay?

>Integer linear programming

can anyone do this?

>can anyone do this?
What have you tried?

Where do you guys get your textbook torrents? Are there any sites other than libgen?
I'm looking for Information Systems Today: Managing the Digital World 7th edition. Can only find 5th and 8th.

I'm not the one yelling for attention in every thread you can. Every regular poster is annoyed at your shit at this point. Here's your (you)

not much, this was an exam question and its pretty hard

>Every regular poster is annoyed at your shit at this point.
I don't mind annoying someone who unironically considers themself a "regular poster".

Usually libgen and bibliotik
your book is not o bib tho

That's why I struggle so hard with physics too.
I can't solve a single damn "hard" problem from those undergrad textbooks out there because I couldn't find the "trick"

What do you guys think about theorems that are almost always true?

For example, I see that Goldbach's "other" conjecture:

projecteuler.net/problem=46

It seems to only have two exceptions (5777 and 5993). Some guy on the problem's forums claims to have checked up to 2*10^13, but I have not seem his work.

It seems interesting that some statements are almost always true, but there are weird exceptions to those rules.

Most physics problems are variations on a familiar theme. I think if you do enough of them (and pay attention to exactly what was done, and find patterns in the solutions that extend beyond any single problem) you'll do fine.

by regular poster I mean anyone who regularly reads, maybe not even posts in, threads. anyone who browses Veeky Forums regularly finds you a tremendously annoying faggot who just has to yell hit catchphrase in every thread you can

We have a joint PMF:
[math]P(X=m, Y=n)=\frac{e^{-7}4^m5^{n-m}}{m!(n-m)!}[/math] for m = 0,1,2,...n and n=0,1,2,3,...
I want to find the marginal PMF for x,y so I do:
[math]P_x(m)=\sum_{n=0}^{\infty}\frac{e^{-7}4^m5^{n-m}}{m!(n-m)!}[/math]
[math]P_y(n)=\sum_{m=0}^{n}\frac{e^{-7}4^m5^{n-m}}{m!(n-m)!}[/math]
I can't see any way to reduce these further, but it just feels like there should be more to this problem. It also asks if X and Y are independent; it doesn't appear to me that these marginal PMFs would become the joint PMF when multiplied. I just feel really weird about this whole problem

>anyone who browses Veeky Forums regularly finds you a tremendously annoying faggot who just has to yell hit catchphrase in every thread you can
Speak for yourself.

>3
Please shut the fuck up you flaming faggots

I'm not a "regular poster", so maybe I don't have the authority to say this, but I think you should take a deep breath.

You are triggered as fuck, bro

Damn, how can the 8th edition be up but no sign of the 7th? shieeeeeeeeeeeeee I'm fugged

...

i checked b-ok.org and unfortunately they only have 3rd edition.

id try torrenting if youre desperate enough (not on uni property tho)

...

What do you mean try torrenting? Don't these use torrents? Is there a torrent aggregator for textbooks? I just need to know if the book is out there before I call it quits.

I have no question I just want to point out that my physics professor is a huge bitch who wouldnt accept my project last semester because he figured out I downloaded an article through sci-hub (sci hub lets you get paywall articles for free). Apparently that ruins the integrity of it. Fuck this guy, Im not paying $35 for an article.

Before you say anything, our uni only has subscriptions to certain journals, not the one I needed

>my physics professor is a huge bitch who wouldnt accept my project last semester because he figured out I downloaded an article through sci-hub
How did he find out?

Hi I found the 7th edition

avxhm.se/ebooks/12920980661292098066.html

Not I have NOT tested the downloads

Barry is a qt but god his handwriting is fucking terrible

If I already have AIDS is there way I can get a "worse" strain AIDS from further use of unprotected sex/dirty needles?

>avxhm.se/ebooks/12920980661292098066.html
Trying it out, you're a gentleman and a scholar.

He repeatedly asked me where I got the article. I was able to kind of bullshit an answer at first but he took me into his office and asked me seriously so I didnt have the balls to lie at that point

Requires payment to download

Nope, downloading atm, just super slow. Unless you mean the pdf file is just an ad with a link. Pls say it ain't so~~

How to [math]x^{x}[/math]
How to solve equation with that term? How to integrate, differentiate it? Just let me know the topics I should study to be able to do it

question for those in physics

when youre reading a text and the author is laying out theory, they might go from an integral to a result or another equation without explaining how. how often do you work out that integral on your own?

Just finished downloading it, THAT'S THE ONE MOTHERFUCKER THANK YOU SO MUCH! How did you find it?

I searched it on that site. Glad I could help

Consider z = x^y. Use elementary calculus. Find the directional derivative along the line x = y. That's for the derivative.

Still thinking about the integral.

I havent done any work on it, but could you consider the random variable x = n-m? Then it looks like the probability density for two indep. Poisson variables. The sum of two Poisson random variables is Poisson, you can do some trick like that.

??? the derivative is super simple, use the exponential

lol, I'm not sure what you mean. Is there something obvious I'm missing?

The derivative is x^x + ln(x) x^x. One part comes from d/dx (a^x) term, and the other part comes from the d/dx (x^a) term. Set a = x.

Can you somehow do this more easily?

Maybe you mean look at the logarithm of both sides, implicit differentiate, etc?

If so, I like this other method. It shows you how it really just reduces to simple rules you already knew.

Let A_1, ..., A_m be subsets of {1, ..., n} such that for all i, j, A_i is not a subset of A_j.

Show that
[eqn]\sum_{i=1}^m \frac{1}{\binom{n}{|A_i|}} \leq 1[/eqn]

Any idea ?
I think I could prove it combinatorially when the sets all have the same size, but that would be ugly and incomplete. There should be a clever solution.

x^x = e^(x ln(x))

jesus im a moron

nah mate it's dumb trivia that one has to see sometime

How do I prove the law of total probability when the universe is partitioned into more than two events? Is it an induction?

it follows from the definition of a probability space, I'm not sure how you would prove it

I'm looking at this one. I hope my progress isn't too trivial so far, but I can prove the case when m < n.

Since 0 < |A_{i}| < n (strict inequality), (n choose |A_{i}|) > n, so that 1/(n choose |A_{i}|) < 1/ n, so since m n, that'll put some kind of bound on the size of the set. You need the size of the sets to be near n/2, or else there is no way to satisfy the condition. The solution feels close!

x^x is by definition exp(x*log(x))

Sorry bro, I'm not really getting anywhere. The case where all sets have the same size is pretty trivial, since there are at most N choose S of them, where S is the size of the sets.

I have an ugly proof if there are two sizes of sets, S_1 and S_2.

My advice would be to think about conditions on the sets if m > n. You should find that they are forced to crowd around n/2, which will force those denominators to get rather large rather quickly.

How bad would it be if some sort of natural disaster prevented people for turning of and completely disassembling one or several nuclear reactors. It takes a while for all fuel to be removed, capsulized and finally buried in an extremely secure and stable underground storage sites where they won't be disturbed for thousands of years. The only reason fukushima and chyrnobyl didn't saturate large amounts of earth with radioactive materials is because people were able to attend to them. How bad would it be if they were left completely unattended and let burn for several weeks, or perhaps even a month? How bad would the fallout be?

That's not exactly the problem. Employers definitely "value" your degree; finding a job in the field without one would be very much harder.
But they also value the degree of the other 150 people that applied for the entry-level position you want.
The fact that people can say with a straight face that 50 job applications is "not trying hard enough" is absolutely absurd but they're not wrong.
No matter how great you did in uni the job market is 100% a lottery without connections anymore and the only way to win is to spam everywhere until they pick you out of the pile for no good reason.

>My friend how majors at a different university downloaded it for me
How difficult is that?
Also, I don't think there is a single professor at my university who will give a fuck where you got the article from.

He wouldnt have bought it. This guy is hardcore, he doesnt take bullshit. Hes some big theory guy, authored multiple textbooks. Plus hes in charge of the graduates so Im not about to risk my future lying. Id rather have to retake a class than get kicked out

Guys, Barry Simon Is A Real Person.

Go ask HIM PERSONALLY, if you actually have 'stupid question' that are on HIS level.

Idk if this helps, but if you take a random permutation of {1,...,n}, the expected number of A_i it preserves is given by the left hand side. This is just linearity of expectation, since, e.g., the probability that a random permutation fixes A_1 is 1/(number of subsets of size |A_1|).

I have no idea why a random permutation should only be expected to fix at most one A_i, but this really feels like a linearity of expectations problem, and that's the easiest way to use it.

user, there are n choose 2 (way more than n) subsets of size 2, and none contains another.

Lads, i want to find all journals about africa ranked by impact factor (or something else relevant). I am clearly not a uni student/professor and don't know how to go about this. google isn't helping much. How do I do it lads?

So is everyone here just a mediocre student social loser? Who would use this site except outcasts?

I Can Verify That Stephen William Hawking Has Been Visiting Veeky Forums.org/sci/ For A Lot Longer Than Simon Troy Cosgrove Has.

... oooh.

That's true, but "way more than n" is relative. N choose N/2 is much larger than N choose 2. If m is on the order of N choose N/2, this suggests you have to use sets of size N/2, which rules out many sets of size 2.

For very large m, the sets must be near N/2.

Or am I not seeing your point?

pretty bad with physics so stuck on this problem which i guess is pretty 101. if you were to close the switch in this circuit, why does it increase current and voltage? i thought it would decrease the voltage since it offers a shorter path to return to the battery. also why does figure 2 in the circuit turn off/become more dim?

Kys FAG

Now THAT, my fellow anons, is 1 True 4channer post here, my kinda pal.

>Underestimated the time I would spend studying
>Now I'm way behind schedule and feel like I'm going to fail all my classes
I study too slow and I'm a brainlet. Please tell me to kill myself.

Tits or GTFO.

My question is where to start, what to read. I know some programming languages, but am new to neural networks so I don't really know what to do.

If you have a 3x3 matrix with 3 independent column vectors and the determinant is nonzero, does that mean the matrix has no solutions?

Hey Veeky Forums, how likely is it for me to learn math up to a calculus level in say 6 months from a geometry level. I need all the help I can get to pull this off. Any good reasorces or videos to watch?

Also I haven't taken a math class in 5 years, is there hope bros? I don't care how hard I have to work, I just want to finish school

Was in the same situation user, a year ago I hadn't been to school (thus taken a math class) in 6 years. Got to Algebra II back then but forgot most of it.

Use Khan Academy to take tests and Google/Youtube to find good tutorials (KA has some good ones but I doubt you'll want to watch 50 3 minute videos a day as opposed to one hour long lecture.)

I'm up to Calc now (though struggling a bit) and its only been a few months. You'll make it user.