« first day (1586 days earlier)      last day (3446 days later) » 

12:00 AM
@GBeau did you do the prod(1-1/a_k) one?
also hi people
 
@anon no I was too busy fluffing my proof for A1 and A2
I felt really good about my A1 though
 
can't remember what those were
 
good luck to both of you
 
A1 was the prime or 1 coefficients
for the taylor series
 
oh yeah that was easy
 
12:03 AM
yeah; I wanted to make sure I didn't miss points on that one
since I knew I had it correct
and A2 was
I can't say "the matrix one" because there were like ~5 matrix problems
 
oh yeah the 1/min(i,j) one
 
yes, that one~
 
you get (-1)^(n-1) over (n-1)!^2 *n?
 
ya
I think there were too good methods
you could expand along the bottom, or you could note the recursion from the reduction
 
I just kept subtracting the first column from the others and reducing the dimensions
 
12:04 AM
yeah
that's how I did it
 
but in retrospect expanding along the bottom would have been easy as well
the start of B was super easy I thought
B1 and B2
B2 I got ln(4/3)
 
yeah
lol @ Patniss and Keeta
 
I think the easiest way to rigorously justify the choice for B2 was changing to a double integral
but it felt so obvious either way
 
@Alizter Any ideas?
 
12:21 AM
a
2
Q: Algebraic function $u\in C^\infty(\Bbb R^n)$

BigMSuppose $u\in C^{\infty}(\mathbb{R})$ is an algebraic function that has no singularities. Can it be said that $u$ is a polynomial? If instead $u\in C^\infty(\Bbb R^n)$ and is an algebraic function. Is it also a polynomial?

 
do you know what he means by algebraic function
 
No idea. I just reworded it.
Because my understanding of algebraic means that the question is a bit silly.
But then again I am tired.
Good night,.
 
12:46 AM
 
a ha! @robjohn is back!
and I'm still messed up on this problem XD
http://math.stackexchange.com/questions/1055016/inverse-fourier-transform-for-the-heat-equation
 
1:11 AM
@GBeau, how 'bout that B1?
Definitely got full points for that.
@anon, I was struggling with the $\displaystyle \prod \left(1 - \frac{1}{a_k} \right)$ one for a while. Might think about it some more when I get back from the gym.
 
1:35 AM
@studentmath are you there?
 
Argh my ears
@Usu yes, why?
 
@studentmath can you help me find the inverse fourier transform of my question?
 
I know absolutely nothing regarding fourier transformations, is it the question you linked above? I will read, but I am probably useless.
 
ya ^
 
If I have any insight I will let you know (but don't count on it)
 
daz
1:38 AM
is anyone familiar with convergence locally in measure?
 
i have a good feeling my bounty will run out without an answer
 
@daz I know the definitions but not much more, why?
 
daz
I would like to know the norm which gives the topology for convergence locally in measure
 
What do you mean?
 
daz
i mean, the norm on the space of measurable (equivalence classes actually) functions for which the metric topology generated by that norm gives precisely convergent sequences which converge locally in meausre
 
1:48 AM
Interesting question, I have no clue - very useless tonight :P
 
Does anyone in here happen to be like an expert at linear algebra?
 
I was...
 
i started learning it a week ago eric!
ask meee
 
but I am trying to finish my pde homework so that comes first
sorry
 
Thats okay usu
 
1:50 AM
@Daz isn't it the standard topology on $L^0$?
 
what is the problem ericc
 
daz
yes
 
Well, its not really a problem but more of a question of whether or not it can be done
 
@EricLawson go
 
daz
i think it has a norm
to be precise
 
1:51 AM
@beginner This question is basically the main thing i need to do asap math.stackexchange.com/questions/1045505/…
 
I think so too, yet I have no clue which it is
 
daz
actually if the measure if finite
 
@beginner I don't know if i can put this into a form that i can use to better the situation, i know that most of it can be put into a determinant form
 
i think that is calculus and not linear algebra since it isn't even close to being linear
 
daz
i have proved that int |f|/(1+|f|) over the whole space
is the right norm
but idk about general or sigma-finite cases
 
1:53 AM
Well, isn't the measure finite anyhow if we are speaking of local convergence?
No wait that's stupid.
 
@beginner Complete bell polynomial can be put into determinant form
 
daz
locally finite
 
Yeah
Doesn't help us though
 
daz
actually for sequences instead of functions, wikipedia gives a norm on little l0
 
@beginner As well as the natural log summation, since it sums up from 1 to n of partial bell polynomials
 
daz
1:54 AM
so it should be similar i would think
 
@beginner The double summation one is a monster though, i don't even know what i can do with that
 
daz
yay found it
 
@Studentmath: It's past your bedtime!
 
1:55 AM
@TedS Had a late-shift work.. I am actually on my way to bed :P How are you? Probability done soon?
 
daz
there is the norm
wonder why it gives the same topology now..
 
68 average on test 3 :( two more classes
 
Ach so. It sounds like an extremely high average here, if it makes it any better
Think that was about the average of those that passed the test in my course :P
 
Cool question for you ... In a 13 card hand, let $X$ be the number of spades, $Y$ number of aces. What's their covariance?
 
@EricLawson i cant help sorry. i tried to take everything non dependant on the riemann sum variable and simplify but i am not experienced enough
@Ted is the 13 card hand random?
 
1:57 AM
Yes @beginner
 
@beginner Its okay
 
I would try to cheat my way using indicators, but I am too tired to think right now
 
does $X$ equal the number of spades you think you will get?
 
You need to use indicators. I even told them to.
 
It's the nubmer of spades you got in that Hand @beginner, considering that hand is random
From a 52 card-desk, right @Ted?
 
2:00 AM
Yup
 
It's a cool one indeed - will go off before Beginner solves it and do it when I wake up :)
 
i probably cant solve it hehe
 
Night, kiddo.
i have been learning probability at age 61, @beginner. It's fun stuff.
 
do i use the formula $\sigma(X,Y) = E[XY]-E[X]E[Y]$
 
yup
but you need a clever way of thinking of $X$ and $Y$
 
daz
2:03 AM
probably some symmetry?
not really thinking about this problem, just taking a wild stab
 
some way that spades and aces interact
 
i dunno @daz
 
cause a spade can be an ace
 
yes, that's important, @beginner.
 
I have a random chain rule question that I can't figure out. Show that on a sphere $\frac{\partial u}{\partial n} = \frac{\partial u}{\partial r}$.
 
daz
2:05 AM
oh i just that problem wrong..
 
not chain rule, @Nick. What's $n$ on the sphere?
 
(x, y. z) / r
 
So what does normal derivative mean?
 
So you get \frac{x}{r} u_{x} + \frac{y}{r} u_{y} + \frac{z}{r} u_{z} but I don't see how that's u_{r}
 
Ok, if you insist, compute $u_r$ by the chain rule.
 
2:07 AM
But doesn't $u_r = u_x x_r + u_y y_r + u_z z_r$ but then $x_r = r / x$ when I calculate it.
I've been staring at this too long...
 
no, you can't take reciprocals like in single variable.
 
hiya
 
Am I ignoring you, @Mike? :)
 
Not that I know
 
hi @mike
 
2:11 AM
Oh? Then how would you work it out then?
 
write out $x,y,z$ in spherical, @Nick, then compute.
 
Ok I'll look at it that way. Thanks!
 
Sure.
 
@Ted What is an event horizon in math words (cf this question)?
 
I imagine it would a solution curve of stability for a system of equations
Similar to a controllable region in optimisation theory
 
2:13 AM
It must have something to do with the light cone. I don't know Lorentzian manifolds.
 
Ok, me neither. I'll avert my eyes.
 
Or not nevermind lol
 
Look at Frankel or Sachs-Wu.
 
No thanks, it was a passing curuosity. :)
 
ignores @Mike after all
 
2:15 AM
I knew that comment would bode ill for me.
 
why did that make him ignored?
 
@KajHansen you can solve for $a_k$ as $2\cos(2^k\theta)$ where $a_0=\cos\theta$ (obviously $\theta$ must then be nonreal) but after that I wasn't sure what to do (besides simplify the denominator of partial products, but not the numerator).
 
Heya @anon
 
heya
 
Soon you get to get your very first evaluations for teaching :)
 
2:17 AM
@mike what did you do wrong?
 
Kaj spent today Putnaming.
@beginner: Don't take me too seriously.
 
ok hehe, i just wanted to not make the same mistake
 
=(
 
Dunno @Mike
 
2:21 AM
I'm just going to wing on this assignment... seriously... I'm pressed for time XD.. I just have to finish 7.2 and attempt 6.4 and viola
 
I've faith.
 
1b and 10 is left for 7.2 whew
anyway is there a fourier transform for $[f(\frac{x+b}{a})]^{\xi}=ae^{ib \xi}\bar {f}(a \xi),a,b$ real, $a \neq 0$ @robjohn @TedShifrin
 
i get $E[X]$ but i dont get $E[XY]$
i cant think of what $E[XY]$ means in english
 
Try Spanish? :)
 
:P
as balarka would say to me if i asked him "go look up the definition, i will not survey you"
 
2:26 AM
You literally multiply the numbers ... And then ask what happens on average.
 
so $E[XY]$ is probabilty they both happen together?
 
where is the fourier transform of $[f(\frac{x+b}{a})]^{\xi}=ae^{ib \xi}\bar {f}(a \xi),a,b$ real, $a \neq 0$? It's not on the table! Does it mean that it's not absolutely integrable?
 
Well, Balarka is far older than you. A whole year. He has to act superior.
 
hehe yeah a year :)
 
For certain random variables, called indicators, yes @beginner
 
2:28 AM
nargh
@TedShifrin is it possible to find a fourier transform of what I just posted?
 
I can't read such complicated math on my iPad ... No longer can do chat jax :(
 
OMG GET ON A PC OR SOMETHING
I'll print screen it and upload it here
 
what level of math is this ted? first year uni?
 
What, the probability?
 
yep
 
2:32 AM
i.stack.imgur.com/h2M1P.png @TedShifrin now can you see? ^_____________^
 
@usukidoll: I am no Fourier transform expert.
 
just tell me if I can take the transform or not for that particular equation
I don't think it's possible because that is not even on the table in the first place
 
@beginner: My course is for juniors and seniors in uni.
tables are not complete.
 
damn it
so it could be possible then
just htf would I take the FT
 
Presumably if the fourier transform of f exists, then that equation holds, so are you asking when a function has a fourier transform @usuki?
if so, your way of phrasing the question is very weird
 
2:35 AM
There aren't tables for this ... You don't have a formula for $f$
 
@anon I got to show that it's absolutely integrable and then I have to use Fourier transform on it, but I'm looking at the FOurier transform table and I don't see it
 
what is "it"?
 
$E[ace]=13 \frac{1}{13}=1$ and $E[spade]=13\frac{1}{4}=\frac{13}{4}$

still working out $E[XY]$
 
Show that if $f(x)$ is absolutely integrable on $(-\infty, \infty)$ then \
1b. $[f(\frac{x+b}{a})]^{\xi}=ae^{ib \xi}\bar {f}(a \xi),a,b$ real, $a \neq 0$\\
@anon
 
You have the conjugate of $f$ on the RHS, not the FT. Right?
 
2:36 AM
@usukidoll see, now that is a properly phrased question!
 
gotta start on the left to get to the right... what else besides $\xi,a,b$ must be real...
@TedShifrin right
 
wat
 
oh yeah and $a \neq 0$ or we're screwed because it will be undefined and there is no absolutely integrable stuff going on.
 
There must be more about $f$ in the early part.
 
yeah but WHAT IS THAT?!
I have to prove that whatever I wrote or took a snapshot is absolutely integrable
meaning it goes from negative infinity to infinity but I have this weirdo equation that doesn't belong in the Fourier Transform table
 
2:38 AM
take logs?
 
@anon how do I tackle this question
 
what else do you know about $f$ from part a?
 
@usukidoll no, you are not trying to prove it's absolutely integrable....
 
so I can disprove it then
 
wat
 
2:39 AM
sdiodfgdsa
omg
 
LOL ... Some impatient people never change. :)
 
the problem gave you the fact that f() was absolutely integrable, did it not?
so you take it for granted.
 
1. Show that if $f(x)$ is absolutely integrable on $(-\infty, \infty)$ then \\


1b. $[f(\frac{x+b}{a})]^{\xi}=ae^{ib \xi}\bar {f}(a \xi),a,b$ real, $a \neq 0$\\

A real of complex-valued function defined on $(-\infty, \infty)$ is said to be absolutely integrable on $(-\infty, \infty)$ if $\int_{-R}^{R} \mid f(x) \mid dx$ exists for all $R>0$ and\\

$\int_{-\infty}^{\infty} \mid f(x) \mid \equiv $$\lim_{R \to\infty} $$ \int_{-R}^{R} \mid f(x) \mid dx < \infty$\\
I took it for granted?
 
the first step to tackling a problem is to understand what it's asking
 
@usukidoll The problem is telling you to take it for granted.
 
2:41 AM
@anon how?
 
Have you never seen an "if X then Y" statement before? You assume X and proceed to prove Y from it.
@TedShifrin WAT
 
I want to take the left side and make it become the right side yay
so do I taket he Fourier Transform or?!
 
Also, this is present in tables for Fourier transforms. See 102 and 104 on Wikipedia.
 
O_O
feels like it's split up
 
@usukidoll That looks pretty ugly
 
2:43 AM
@usukidoll to compute the fourier transform of f((x+b)/a), write down what that means, then manipulate it to get what you want
 
@robjohn ikr.........my book didn't have the FT of it so I'm like wth... I am just going to take the FOurier Transform of the left to get to the right... and by the table from the right to the left
@anon using the combination of 102 and 104 on wikipedia right?
 
yes
 
which means I have a $f(x-a)$?! How is that possible ?
 
if you wanted to, you could use the rules one after the other to go from f(ax+b) to what you want
 
@usukidoll There is an $x$ on the left hand side but not on the right.
 
2:46 AM
@robjohn... yeah, I think starting from the right would be much easier
 
you shouldn't be trying to use tables anyway; you should be proving it from scratch probably, just using manipulation of integrals
@robjohn x is a dummy variable on the left
 
ughhhhhhhhhhhhhhh
I'm starting on the right to get the left..through those tables
hehehe
 
stop using tables
 
=(
 
@anon but it should have a counterpart on the right, no?
 
2:47 AM
turns the tables on @anon
 
@robjohn does $\int_0^1 xdx=1/2$ have an $x$ on the right?
 
$E[XY]=\frac{13}{52}$ chances that you get the ace of spades at all, plus the chance you get an ace or a spade at all, but not both cause you did that first, so the ace is already done and that spade is gone as well: $\frac{13*12}{52}$
so we get $E[XY]=\frac{169}{52}=\frac{13}{4}$ and $E[X]E[Y]=1*\frac{13}{4}$ so we $\sigma(X,Y)=0$?? woops typo
 
:/ how is this even possible?!?!!??!!?
 
@anon But that $x$ is bound to the integral, the $x$ in the other statement is not bound, as far as I can see.
 
5 mins ago, by anon
@usukidoll to compute the fourier transform of f((x+b)/a), write down what that means, then manipulate it to get what you want
 
2:49 AM
which means I have to somehow find $f(\frac{1}{a})$ and f(x+b)
 
@robjohn the other statement is an integral, just not in any standard notation. You're taking the fourier transform of the function f((x+b)/a), which will yield a function of \xi, no x at the end.
 
maybe $f(\frac{1}{a})f(x+b)$
 
@usukidoll no
 
-.-
 
do what I told you to do
 
2:49 AM
Perhaps I am misunderstanding what $\left[f\left(\frac{x+b}{a}\right)\right]^\xi$ means
 
is that right @ted $0=\sigma(X,Y)$
 
@robjohn it is ${\cal F}_x\{f(\frac{x+b}{a})\}(\xi)$, as I might write it
@usukidoll How about this. What does "the fourier transform of f(x)" mean? Write down the definition.
 
using

Fourier Transform

$f(\frac{x+b}{a})e^{-i \xi x}$
 
anyone know if there's a round-bracket ewuivalent of llbracket? ie something that packs (( close together for writing the field of formal Laurent series
 
oh wait i think i did it wrong i see woops, ill redo it
 
2:51 AM
I didn't even know about llbracket, heh
 
$f(\frac{x+b}{a})e^{-i \xi x} \frac{1}{\sqrt{2 \pi}}$
nargh
 
@PedroTamaroff I've plenty of time to read your notes now. The Theorem in section 3.2 is false: the rhs of the iff is the same as saying every field that is a finitely generated $A$-algebra is integral over $A$, implying $A$ is a field. But, obviously there are Hilbert rings that are not fields.
 
@anon Oh... that makes some sense then.
 
wait no it is right
 
@beginner:
 
2:52 AM
yeah no covariance
 
it's in a special package, stmaryd, which I use only for that
 
i wish more of my students had your enthusiasm ;)
 
but stmaryrd doesn't have a round-bracket ewuivalent
 
is it right?
 
@PedroTamaroff Since I know the rabinowitch trick and whatnot, I am pretty much done with reading your notes btw. I can give some ideas for changing the layout of the notes and some perhaps interesting ways to expand on what you have via email soon.
 
2:53 AM
Fourier Transform

$\int^{-\infty}_{\infty}f(x)e^{i \xi x} \frac{1}{\sqrt{2 \pi}}$
 
is it wrong noooo
 
Yes, but I'm not convinced you have a valid solution.
 
stupid phone doesn't know the word equivalent
 
maybe I should let $d = \frac{x+b}{a}$ and then go from there or something ugh
 
is my $E[XY]$ wrong?
 
2:54 AM
$\int^{-\infty}_{\infty}f(\frac{x+b}{a})e^{i \xi x} \frac{1}{\sqrt{2 \pi}}$
 
@usukidoll you forgot the dx, and I'd put the normalizing constant on the left, but otherwise good. now replace f(x) in that expression with $f(\frac{x+b}{a})$ and do some substituting.
 
$\int^{-\infty}_{\infty}f(\frac{x+b}{a})e^{i \xi x} \frac{1}{\sqrt{2 \pi}} dx$
 
What is your $E[XY]$?
 
$\frac{1}{\sqrt{2 \pi}} \int^{-\infty}_{\infty}f(\frac{x+b}{a})e^{i \xi x} dx$
there we go
now how do we subsitute this crap?
 
@usukidoll perfect. now what happens when you apply the substitution $u=x+b$, to start off with?
 
2:55 AM
OH!
 
after that you'll do the substitution $v=u/a$...
 
@ted $13$ chances to get the ace of spades in the 13 cards. + chances to get spade of not ace so 13 chances to get one of the 12 cards(spade not ace) in the 52, so $E[XY]=\frac{13*12}{52}+\frac{13}{52}=\frac{13}{4}$
 
$\frac{1}{\sqrt{2 \pi}} \int^{-\infty}_{\infty}f(\frac{x+b}{a})e^{i \xi x} dx$
Let u = x+b

$\frac{1}{\sqrt{2 \pi}} \int^{-\infty}_{\infty}f(\frac{u}{a})e^{i \xi x} dx$.

Now if we let $ v = u/a$ then .

$\frac{1}{\sqrt{2 \pi}} \int^{-\infty}_{\infty}f(v)e^{i \xi x} dx$.
 
You're rewriting in terms of indicator functions without knowing it ...
 
so it is right but i did it the wrong way?
or i solved it by fluke
 
2:59 AM
ok... so I need to find the fourier transform of $f(v)$ @anon
 
what are fourier transforms for usuki?
 

« first day (1586 days earlier)      last day (3446 days later) »