« first day (2051 days earlier)      last day (3267 days later) » 

00:00
Equation 1 says that we can pull in the integral sign on your RHS
And in Equation 2, we are just looking at the second component of the brackets, for both sides of your equation
Don't we have < v(y), ρ(x-y) > ?
What property do we use to get this?
I don't know
Because I don't know what your integral sign means
From $\langle u(y), \int \phi(x+y) \rho(x) dx \rangle$
00:03
It's certainly not a Lebesgue integral
So I just took it as a definition: $\int T(x) f(x) dx := T(f)$.
(T a distribution, f a test function)
Isn't it the inner product?
Sure, yes, I'm also using $\langle T, f \rangle = T(f)$.
Why?
Because nothing else makes sense!
Aha
00:06
:P
But with that definition we do not get the above integral
do we?
Also this is actually fairly common notation: if $L$ is an element of $V^*$ and $g$ is an element of $V$, then $\langle L, g \rangle := L(g)$.
Sure we do.
How?
Definition of convolution
Oh I see
You are doing too many steps at once
I am only talking about Equation 1
Which one is this?
00:10
in your document
I'm not going to rewrite it
the first thing in displaymode
Oh, it's numbered
convenient
(5.2.2)
Hi Rob John, I have a question about something I'm trying to do with what you showed in this post: math.stackexchange.com/questions/270566/…
I'm trying to use the formula for the fourier transform of e^-n^2(x-m)^2 you proved to get a kernel for the heat equation with lower order terms
Some of my work is documented here: math.stackexchange.com/questions/1699440/…. I was wondering if you had any thoughts on what might be going on. Thank you!
You mean that we use the fact that (5.2.2) is equal to $\langle \phi(x), \langle u(y), \rho(x-y) \rangle \rangle$, right?

And how do we deduce that this is equal to $\langle \rho(x) \otimes u(y), \phi(x+y) \rangle$ ?
Hello everyone. Do you guys have any idea how to solve for F in the matrix equation $F^T F = M$ where T stands for transpose..
00:30
@EricStucky Did you see my question?
00:45
@EricStucky We have that $|G|=p^km$, where $k\not\mid m$. That's why a $p$-Sylow group is a group of maximum order $p^k$, right?

Since $N_G(P)\leq G$, we have that $|N_G(P)|\mid |G|$. Therefore, the prime factorization of $N_G(P)$ has no more than $k$ factors of $p$.

We have that $P\leq N_G(P)$, so $|P|\mid |N_G(P)|$. Therefore, $N_G(P)=p^kn$, where $k\not\mid n$ and so $P$ is a $p$-Slyow subgroup also of $N_G(P)$, right?
01:18
@r9m And there is a simple pole at the origin.
 
2 hours later…
03:26
hi
Is the Hilbert space $\ell _2$ the same as $\mathbb R ^\omega$ with the product topology?
is the product of two functions also a function?
@ForeverMozart
do you know what is degree of isogeny ?
r9m
r9m
@RandomVariable cool man! That's far more elegant! :D Thanks!
$\omega$ is the first countable Cardinal?
03:42
@MikeMiller do you know what is the meaning of "degree of isogeny"
@MikeMiller do you know the answer?
I think it is true
apparently it's not trivial, proved in 1966
I'm surprised it's true.
03:57
yes it is interesting
I assume they are putting the product topology and not box topology on $(0,1)^\omega$, because they mention it is a natural subspace of the Hilbert cube (which has the product topology)
@MikeMiller
So $F_2$, the free group on two generators, has normal subgroups isomorphic to $F_n$ for $n\ge2$
I believe that $F_3$ has normal subgroups isomorphic to $F_n$ for $n\ge3$ and odd (so, no normal subgroup isomorphic to $F_4$). Is this correct?
@AkivaWeinberger: Wrong question. The correct question is about finite index normal subgroups.
In which case, correct.
Oh, I see
Finite index means finitely generated, yes? EDIT: No
If I remove the finite index constraint, but I add the fact that $F_{\aleph_0}$ can also be a normal subgroup, is it correct?
No, I don't believe so.
I think it's probably true that $F_4$ (and $F_2$) are normal subgroups of infinite index. I'm not actively thinking about it, but I'm rather certain it's true.
Are parametric equations specific to rectangular coordinates? As far as I've seen, a parametric representation of a curve in $\mathbb{R}^2$ is often characterized as $x=f(t),y=g(t)$ where apparently $x,y$ are assumed to define points in the Cartesian $xy$-plane.
That is, why not for example allow $r=f(t),\theta=g(t)$ for a polar coordinate system? Surely the two variables can be parametrized as functions of $t$. I'm likely missing some obvious caveat.
There's not really any caveat nor reason not to do that. It's maybe slightly more annoying to talk about smoothness of curves thag pass through the origin in polar.
Thanks. Hmm... I guess it's a lack of usefulness / application.
And the peculiarities associated with some non-rectangular systems.
04:48
@Evinda: I feel we're running around in circles on this one. The short version is that there is no cleverness in this theorem; it's all definitions and possibly linearity of distributions. Equation 1 (i.e. 5.2.2) follows by definitions, Equation 2 is a change of variables, and Equation 3 combines the two. Do you understand 5.2.2?
@MaryStar: Yes, and yes.
Nihao
What are the subobjects of a Kleisli category?
05:22
@AkivaWeinberger The way to think about these things is write down the induced covering spaces on graphs.
05:56
God this wikipedia article is awful. en.wikipedia.org/wiki/Stein_manifold I always get this uneasy feeling when theres so many egregious errors in something I know a little about, and feel like I must be missing all these errors in things I don't know anything about.
r9m
r9m
06:16
@robjohn did you finish your writeup? :)
06:39
@r9m yes.
07:17
$\lceil$
Got it.
Had to test some latex in here because the plugin broke while editing one of my really long answers.
r9m
r9m
@robjohn cool, I am posting the question in main now :-) Thanks!
@r9m answered
r9m
r9m
@robjohn ooh! yea .. thanks! lemme read :D
@robjohn (+1) :D Thanks! Very clear and nice presentation!!
says I have to wait 11 mins before I can accept an answer! Never had a question answered this fast before :D
07:35
@r9m We can convert the Harmonic numbers to $\psi$ functions
r9m
r9m
@robjohn yup! got that! :-) There'd be a possible application of reflection formula in there too :)
@r9m the signs are wrong, I believe.
$H(x)=\gamma+\psi(1+x)$
07:54
@robjohn,can you help me?
@geromty with?
see this link
08:15
@geromty Will Jagy's answer looks good to me.
How would I define the subobjects of a free monoid object in some category $\mathcal{C}$?
@JulianRachman Hmm, so your category has a product and is also concrete?
Ya
(and as a side note, are you familiar with Kleisli categories?)
Not at all, no
Oh ok. Carry on
08:21
Hmm, not actually sure which is a good way to define subobjects then. Might depends on what you need to do with them
Like take $A$ to be a singleton $\{1\}$. Then $A^*$ is the natural numbers
so what would the subobjects of the natural numbers be and how would you define them?
I want to prove that this category is noetherian.
@JulianRachman I thought you meant to start with an object in a category which was a free monoid object in that category
@JulianRachman Which operation is $A^*$?
what are the morphisms in the category $A^*$?
Concatenation @Axoren
08:24
Kleene Star?
Yep
@Tobias they are just morphisms between free monoids
It wouldn't be all of the natural numbers unless you were in base 1 (the tally system)
I guess you can say that you have a category of free monoids
Why do you say that @DeNiSkA ? Looks like a correct proof to me. Assuming that $[x]$ has the IMO default meaning: $[x]=n$ for all $x\in [n,n+1)$ for all reals $x$ and all integers $n$. — Jyrki Lahtonen ♦ 9 mins ago
@Axor Ok. But then how would you still define the subobjects of what you have described?
08:25
@JulianRachman the subobjects of the naturals should be free submonoids
Example?
And lets all get on the same page
my appologies
@JulianRachman multiples of any fixed natural number
ok we are now going to take the concatenation A^* where {1}=A then {1,11,111,1111,...}=A^*.
Now tell me what the subobjects of A^* are
Are we taking the natural numbers to be von Neumann integers, then each natural number is a set containing all natural numbers less than it.
@Axoren Why on earth would that be relevant here?
"what the subobjects of A^*"
@JulianRachman in the category of free monoids, the subobjects are precisely those consisting of multiples of a fixed natural number
@Axoren in terms of category theory, not sets, obviously
but you have the wrong A^* @Axoren
I guess I misinterpreted what you meant by "But then how would you still define the subobjects of what you have described?" @JulianRachman
But we are talking concatenation here @Tobias
08:29
In your set, you have a different set.
A^*={1,11,111,...}
@JulianRachman Yes, I know, I am just identifying with the naturals under addition as that is easier
@Axoren This is completely unrelated to what sets are used to represent things
Ok. But actually I need the concatenation one.
@JulianRachman you should easily be able to translate back
08:32
The free monoid in one generator can be described in either of the two ways, and translating between them is easy enough.
So subobjects of the concatenation A^* would be {1}, {11}, {111}, ...?
no, those are not subobjects (unless you now mean the brackets to mean "generated by")
Hello, I am new here.
@Algebra2015 Hello New Here
I applogize in advance for asking questions
08:34
@Tobias Ok you are going to have to walk me through
I have one request
@JulianRachman The natural number $n$ corresponds to the string with $n$ $1$'s
Can I isk here a direct question ?
@Algebra2015 yes
or how does it work here?
08:35
@Tobias Ok. I follow
@Algebra2015 You've asked one just now. No one's bitten your head off yet :P
are here also "private-chat" help "rooms"?
:)
no offense, please. Im fresh
I'm fairly certain you can invite users to chatrooms, but they're not private I don't think.
@Algebra2015 It is possible to make a new chat mainly for specific people (I don't think you can keep others out though), but this is not used so much
You may need higher reputation to create one.
08:36
ok
there are also a few specific rooms for specific topics, but mainly people use this one (it rarely gets too crowded)
The natural number $n$ corresponds to the string with $n$ $1$'s - Anything more to add? @Tobias
@JulianRachman Not really, that should help you see what the subobjects are given y description of them inside the naturals
(note that this is in the category of free monoids. There are more subobjects in the category of monoids)
Can you give me an example OF a subobject in this category?
@Algebra2015 You never did ask your direct question, unless the question was about the chatrooms.
08:38
@JulianRachman All strings of even length
How to solve this ? Z-Z^T=BC^{-1}; \$C=\pmatrix{1 & -1 \\ -1 & 2}$\
@Tobias so what if we consider the power set of subobjects?
Is that considered the collection of all subobjects?
Surround latex in \$ \latex \$ like so: $\sum f(x)$
@JulianRachman Not sure what you mean by powerset (that usually means all subsets)
08:40
This problem, right? $Z-Z^T=BC^{-1}$, $C=\pmatrix{1 & -1 \\ -1 & 2}$ ?
@JulianRachman Yes, but the term does not fit in the way you used it here. Do you just mean the set of subobjects, or do you mean the powerset of the set of subobjects?
You may need to use the latex plugin found here to see it: math.ucla.edu/~robjohn/math/mathjax.html
i CLICKED ON rendering on
set (collection) of subobjects
@Axoren , YES
08:42
@Algebra2015 What is being solved for here, Z or B? Is there any additional information?
@JulianRachman I am not sure what you mean. You seem to be asking whether the set of subobjects would be considered the collection of all subobjects (yes, by definition)
Yes your assumption is right
@Algebra2015 What is Z, then?
Z=\$pmatrix{a & b \\ c & d}\$
08:45
So then how would you define your assumption in terms of category theory?
@Algebra2015 Then, if you know that, you could tell me what $(Z - Z^T)$ is, right?
@JulianRachman What assumption?
"You seem to be asking whether the set of subobjects would be considered the collection of all subobjects (yes, by definition)"
@Axoren. a-c, b-d, c-a, d-b ?
@JulianRachman The set of X always means "the collection of all X".
08:47
@Algebra2015 Not quite. What is $Z^T$?
It's the transpose, but what's the matrix?
@axoren not to "junk" here ? can we go in "private"
Well I want to prove that the category is noetherian
how would I do this?
@Algebra2015 There's no junk here. Quite often there are numerous mathematical discussions overlayed on eachother.
08:48
You're not being a bother here.
ok
didn know
It's fine. Don't worry.
@JulianRachman Which category? The category of free monoids?
@Algebra2015 Now, do you know what transposing a matrix does to it?
08:50
$Z^T$ = 1st row: a c, 2nd row: b,d
Yes. So, $a$ and $d$ haven't changed positions, correct?
@JulianRachman But that category has non-noetherian objects, so it cannot be noetherian (the free monoid on infinitely many generators)
@Axoren. ok and than ?
then
@Algebra2015 Now try again and tell me what $Z - Z^T$ is.
Before, you said that the top-left corner was $a-c$.
That's incorrect.
@JulianRachman what does that mean?
08:52
What if we have finitely many generators instead?
0, b-c, c-b, 0
hmm, that is more tricky (since this not only rules out some objects but also a lot of the subobjects we would have had before). No idea how one would do it
Ok new situation:
@Algebra2015 That's correct, so now you have this new matrix on the left-hand side of the equation. Let's call it $Z'$ or something. $Z' = \pmatrix{0 & b-c \\ c -b & 0}$
Now you should have $Z' = BC^{-1}$ and you know what $C$ is.
... then
08:55
So, how can you isolate $B$ by itself on the right-hand side?
If I define the subword order on objects of the category of preorderd monoids, how would I prove that this order is well-quasi-order?
Hint: $A^{-1}A = I$
@JulianRachman No idea
Then do you know of an example of a noetherian category?
@JulianRachman sure, take any finite category
08:57
ok
Well wait
we can make A^* finite
by taking $\coprod_{k\geq0} A^k$ where
@Algebra2015 What can you do to both sides of the equation to leave $B$ by itself?
I dont know :(
we take the concatenation up to $k$-length strings over $A$.
I gave you a hint. $A^{-1}A = I$
So, if $Z' = BC^{-1}$, what happens if you multiply both sides of the equation by $C$ on the right?
09:03
z'*c=b ?
Yes.
@Tobias ^
$Z'C = B$
You know $Z'$ and you know $C$, so you can multiply them together to get $B$
Remember $Z' = \pmatrix{0 & b-c \\ c -b & 0}$
$C=\pmatrix{1 & -1 \\ -1 & 2}$
$Z' = BC^{-1}$
$Z'C = B$
$Z'C = \pmatrix{0 & b-c \\ c -b & 0}\pmatrix{1 & -1 \\ -1 & 2} = B$
09:05
I meesed with letters
how is it seen in matrix form
You good? @TobiasKildetoft
solution for B matrix is {1 & -1 \\ -1 & 2} = B$
?
You should be able to multiply those out yourself, @Algebra2015.
No, $B$ is equal to the product.
That's all one equation
You know Matrix Multiplication, correct?
SOLUTION is {-(b-c) & 2*(b-c) \\ b-c & -(b-c)} ?
Almost. The bottom row is wrong.
It should be $c-b$, $b-c$
09:10
what represent rows her: {1 & -1 \\ -1 & 2}
Rows are separated by \\
a & b \\ c & d $\to \pmatrix{a & b \\ c & d}$
i multiplied "your" 2 matrices: \pmatrix{0 & b-c \\ c -b & 0}\pmatrix{1 & -1 \\ -1 & 2}
:)
and I got: {-(b-c) & 2*(b-c) \\ b-c & -(b-c)}
The bottom left corner should be (1)(c-b) + (-1)(0)
09:12
aaaaa. yes
sorry
The bottom right corner should be (-1)(c-b) + (2)(0)
It's alright, just be a bit more wary.
i misstyped; solution is then: {-(b-c) & 2*(b-c) \\ c-b & (b-c)} ?
How many solution exist?
in overall
?
@TobiasKildetoft Let me know if your working on the problem or not.
@Algebra2015 Are they any values $a, b, c, d$ that you can't pick for $Z = \pmatrix{a & b \\ c& d}$?
Note, the answer doesn't care what $a$ and $d$ are. They can be anything
NO. Just wanna know are there endless solution if a,b,c and are real numbers?
:)
09:19
@Algebra2015 If there aren't any numbers you can't pick, then you can pick all real numbers.
so the solutin can be formally written: a,b,c and the are reAL numbers with solution for B matrix {-(b-c) & 2*(b-c) \\ c-b & (b-c)} ?
Yup.
tnx :)
queation: are here on SE also chats for e.g. chemistry, computer science?
@axoren
:)
Yes, but those are under different sites. Computer science is likely stackoverflow.com
Unless you're asking more Computer Science theory questions
and physics?
chat
tnx
tnx
Can u help end this task ?
It's actually fairly late where I am
I'm about to log off and go to bed.
Good luck
5 more minutes?
please
tnx
09:51
@EricStucky Ok, I got it... The theorem (4.1.1) is the following:
How do we deduce from this that $ \langle u(y), \rho(x-y) \rangle \in C^{\infty}(\mathbb{R}^n) $ ? @EricStucky
@JulianRachman I am not (I had to go to a meeting)
Hey @robjohn
@robjohn Do you maybe have an idea how we deduce from the above theorem that $ \rho \ast u= \langle u(y), \rho(x-y) \rangle \in C^{\infty}(\mathbb{R}^n)$ ?
What is a final formal math write-down for a solution of: math.stackexchange.com/questions/1696842/…
mathmaticians are strict
10:50
hi... Does anyone have a feeling if the following could plausibly be true? Consider a real valued positive definite matrix $M$. Is the following inequality true?

$$\sum_{k\in\mathbb{Z}^n} e^{-k^T M k} \leq \prod_{i=1}^n \left( \sum_{x \in \mathbb{Z}^n} e^{-\lambda_i x^2} \right) \ .$$
 
1 hour later…
r9m
r9m
12:05
@robjohn Finally found a video for your elegant proof of the integral: Ex-wife (cut to 0:56 onwards)
12:17
@Evinda It looks as if it is a direct application. Where are you getting stuck?
@r9m does Chris'ssis still come around ? :(
r9m
r9m
12:34
@Hippalectryon haven't seen her in some time .. I guess she is busy.
@r9m I'm just hoping she doesn't forget to notify us when her book is available :-) she doesn't have our (well at least my) email
r9m
r9m
@Hippalectryon ah! okay ,., but I think she will come here and announce the good news when her book is published
I hope so :DD
@Hippalectryon It's not like we are hard to find.
@robjohn I've been a lot less active here lately though
r9m
r9m
12:52
@robjohn can we extend the method to $\displaystyle \int_0^{\pi/2} \left(\frac{\sin ax}{\sin x}\right)^n\,dx$ too? :-)
@r9m Possibly, but I'd have to look closer.
r9m
r9m
@robjohn 'kay! :) Well in that case this method could be used to get $\displaystyle \int_0^{\pi/2} \left(\frac{x}{\sin x}\right)^n\,dx$ as well, dividing by $a^n$ and letting $a \to 0$ ..
@robjohn it was asked here back in '14
Hey guys question on some topology -
Let $Y$ be the space obtained by gluing two points $D = \{a,b\}$ on a torus $T$ together, $Y = T/D$
trying to see how we obtain the exact sequence
$0 \to \tilde{H^1}(D) \to H^1(Y) \to H^1(T) \to 0 $
sorry i mean
$0 \to \tilde{H^0}(D) \to H^1(Y) \to H^1(T) \to 0 $
I think i see
do we have an exact sequence of chain complexes $0 \to C^\bullet(D) \to C^\bullet(T) \to C^\bullet(Y) \to 0$ ?
@iwriteonbananas you do topology :)
13:12
Sure, sometimes...
Your exact sequence should be $\cdots H_i(D)\to H_i(T)\to H_i(Y)\to\cdots$
isn't there an exact sequence in cohomology as well?
Oh, for some reason I thought you were doing homology
no sorry
In that case your sequence should be
$\cdots H^1(Y)\to H^1(T)\to H^1(D)\cdots$
and that follows from the short exact sequence of chain complexes I wrote right?
13:15
Sure, or alternatively it's easy to derive from the ordinary LES in cohomology
yep
now what about that reduced bit o_O
I think i see that as well
so if we start with the LES in homology
Right, just look at the reduced cohomology LES of the pair $(T,D)$: $$0\to \tilde{H}^0(D)\to H^1(T,D)\to H^1(T)\to H^1(D)\to\cdots$$ and use the fact that the quotient map $(T,D)\to (Y,pt)$ induces an iso $H^*(T,D)\cong \tilde{H}^*(Y)$
yeah
thanks!
13:55
hi @JC574 do you know what is isogeny ?
Is the product of regular (invertible) matrices invertible?
hi... $x^TMx \geq \lambda_{min} x^Tx$.. is that right?
morning chat
14:12
morning@Semiclassical
Though it is night here
details!
@r9m I don't know that a general formula for all $n$ would necessarily be possible, but for some $n$ other than $2$, it can probably be done.
@robjohn what I'd be curious about myself: I can view that integral as (half of) the zeroth coefficient of the Fourier series of $\sin(ax)/\sin(x)$ on $(-\pi/2,\pi/2]$
can one obtain the full fourier series?
the relation to the case of $n>1$ is then that the Fourier series is determined in terms of the $n=1$ case
14:48
@L33ter I haven't seen it before but wikipedia defines it
looks like a surjective morphism of algebraic groups with a finite kernel
Random category-theory question: What's the difference between a monoidal category and a tensor category? The latter doesn't seem to be consistent in what i've seen via google.

« first day (2051 days earlier)      last day (3267 days later) »