« first day (2063 days earlier)      last day (3254 days later) » 

04:00
why is Cn(B)/Cn(A n B) --> Cn(A + B)/Cn(A) is an isomorphism ?
I learned them in [very] different contexts, my ears are perked :)
@BalarkaSen I'm sorry, say again?
The smash product is actually a some sort of dual of the cup product, if you will. More specifically, if $H^n(X; G)$ is the $n$ cohomology group of $X$ with coefficients in $G$, then $H^n(X; G)$ is isomorphic to $[X, K(G, n)]$ (collection of homotopy classes of maps $X \to K(G, n)$ - which is actually a group because $K(G, n)$ is a homotopy commutative homotopy associative $H$-space).
And the cup product on cohomology can be realized as maps to smash of $K(G, n)$'s.
@Adeek Hatcher explains: it's a bijection at the level of basis of those two free groups.
what do you mean ?
@EricStucky I think the keyword is "spectra" for these kind of stuff. I don't really know anything other than what I said above.
04:04
can you explain more ?
haha oh wow, this is a much deeper rabbit hole than I was expecting
I have heard the word 'spectra' before :D
me 2
Sounds like a comic book character
I went to couple of topology talks where people used spectral sequence
to do some shit
@Adeek $C_n(B)/C_n(A \cap B)$ is a free abelian group. What's it's basis?
@AkivaWeinberger What do you want me to say again? :)
04:07
@EricStucky: Well of course you have, every undergrad learns about the spectrum of the Laplacian
it is the basis elements of B ?
XD
although wait
what?
so it is the basis elements that don't lie in A
I took a course in spectral graph theory so I saw that, but I don't think this is standard material o.O
right ?
04:08
@Adeek It is not true that image of basis elements of $B$ by the quotient map $C_n(B) \to C_n(B)/C_n(A \cap B)$ gives a basis.
because since we are quotiening by
Cn(A n B)
Yes, you want singular simplices not sitting in $A$.
oh right the Laplacian is a calculus thing ^.–
yes
What's a basis for $C_n(A + B)/C_n(A)$?
04:09
it is the basis elements of Cn(X) that isn't sitting inside of A
I get it
oh I see
good I understand
so it must sit inside of B
@Adeek basis elements of *Cn(B).
and not A
yeah
oke oke good
thanks a lot blarka
You just have to check that the map induced from inclusion is a bijection on basis.
Then it's an isomorphism on the free abelian groups generated by them.
yeah which it is by the information before from the D maps
yeah I understand
1 second I am gonna go home from library I will doing more topology
until like 5 am
brb
@EricStucky Yes, so smash product apparently is useful in homotopy theory. At least I have seen it cropping up everywhere "stable homotopy theory" is mentioned.
I don't know any of this stuff, of course.
04:14
I was joking.
Probably most undergrads don't see the word Laplacian, and most also don't see the word spectrum
I would probably replace spectrum of a Laplacian by spectrum of a ring.
That'd make it more hilarious as it'd be only applicable for French undergrads.
well that's a relief :P
lol
mainly because most don't learn math in english
It wasn't funny, so most modifications would make it funnier.
Oh man I love these things
"Pick up the book closest to you and turn to page 45. The first sentence explains your love life."
"If $\epsilon>0$, Step V shows that there are sets $K_i$ and $V_i$ such that $K_1\subset A\subset V_1$, $K_2\subset B\subset V_2$, and $\mu(V_i\minus K_i)<\epsilon$, for $i=1,2$."
04:27
$d_{\infty}(\vec{x},\vec{y}) = \sup_k |x_k - y_k|$
"Let $I_s$ be the irrelevant ideal inside $k[S]$ consisting of polynomials with terms of degree $\geq s$". This defines the empty set in $\Bbb P^n$"
@EricStucky Sorry but that's garbage.
Yours is better :P
What are example balls around that metric?
04:29
Closest non-math book: "No olvides que son refugios de emergencia."
Which translates to "Don't forget that they're emergency shelters."
lol wut
Um, so my love life will be a Spanish emergency shelter?
Or maybe there would be a lot of emergency shelters in your love life.
anyone?
$d_{\infty}(\vec{x},\vec{y}) = \sup_k |x_k - y_k|$. What are example balls around this metric?
@JesterTran Squares with sides parallel to the axes, I think
Squares/cubes/hypercubes
04:32
@AkivaWeinberger thanks
(I have The Martian in Spanish. I haven't finished it because apparently I don't actually know Spanish)
(El marciano)
Actually maybe it was not garbage, @EricStucky. One can perhaps interpret it in the way that my love life is an empty set, thus by Nullstellensatz cut out by things from the irrelevant ideal.
Oh, you were referring to your quote? It might not be accurate, but 'Let $I_s be the irrelevant ideal' is a pretty applicable fragment.
Really? How so?
04:39
'Irrelevant' and 'ideal' have other meanings :P
I have no idea what you're talking about, but I believe you.
@AkivaWeinberger how do we visualise this metric?
How would you visualize the standard metric, Jester?
@EricStucky I'm really confused now and it's hard to go back to basics
what is a standard metric?
$d(x,y) = \sqrt{ (x_1-y_1)^2 + \cdots + (x_n-y_n)^2 }$.
04:44
@EricStucky right
distance between two points
Okay but that's not really a visualization :P
What I was getting at with this question is: I don't think of a metric as something that is 'visualizable'; it's a method for measuring "distance" (in an abstract sense) between two points.
@EricStucky so how did the other person successfully find a ball around the infinity metric?
Oh, you can visualize those sometimes.
But you just have to work through the definitions
in $\Bbb R^2$, for instance.
the metric between two points is just, take the difference in the coordinates, and pick the bigger one.
So you try to think about which points are at distance 1 from the origin.
And you get out your pencil and you draw something.
a square?
yep :)
04:49
wait
ello skull :)
yep yep
I'm stuck on an assignment question which asks to find the boundary and interior, due tomorrow but I really need to do this alone
05:07
best of luck
thanks
 
3 hours later…
08:22
hey everyone
can I access online Matlab by my Moodle account?
 
2 hours later…
10:01
Has this identity appeared on MSE before?
10:16
Hello!!
Is the group $\mathbb{Z}_2\times\mathbb{Z}_2$ isomorphic to the dihedral group $D_2$ ?
10:43
Is $D_2$ cyclic?
No, since it is generated by two elements, or not? @Krijn

$D_2=\langle a, s \mid a^2=1=s^2, asa=s^{-1}=s\rangle$
So it is not isomorphic to $\mathbb{Z}_4$
And it has 4 elements
So yes, it is isomorphic to $\mathbb Z_2 \times \mathbb Z_2$
Although you could have also just written down an isomorphism
So, are all the groups with $4$ elements isomorphic to either $\mathbb{Z}_4$ or $\mathbb Z_2 \times \mathbb Z_2$ ? @Krijn
@Krijn You mean to find a map from $\mathbb Z_2 \times \mathbb Z_2$ to $D_2$ that is bijective and an homomorphism, right?
Yes, to both
We have that $\mathbb Z_2 \times \mathbb Z_2=\{(0,0), (0,1), (1,0), (1,1)$ and $D_2=\{e, a, s, as\}$, right?

Could we define the following map:
$h:\mathbb Z_2 \times \mathbb Z_2\rightarrow D_2$ with
$(0,0) \mapsto e \\ (0,1) \mapsto a \\ (1,0) \mapsto s \\ (1,1) \mapsto as$
? @Krijn
10:53
Yes that seems okay
Almost any sensible map will do actually
As long as $(0,0) \mapsto e$
Try to find out why
Since each element of $\mathbb Z_2 \times \mathbb Z_2$ is mapped to one element of $D_2$, and since for each element of $D_2$ there is one element of $\mathbb Z_2 \times \mathbb Z_2$ that is mapped to that, we conclude that themapping is bijective, right? Or is the justification wrong?
We have that each element of order $4$ is abelian, right?
To show that the mapping is an homomorphism do we have to show that $h_1\circ h_2 (x)=(h_1\circ h_2)(x)$, or not? @Krijn
Yes to the first and last question, but I don't understand your second question?
What do you mean when an element is Abelian?
Furthermore, no element is of order 4
Sorry, I meant a group of order $4$ is abelian. @Krijn
Ah, yes, then it is true
11:15
@BalarkaSen hey do you have a little bit of time?
11:34
Sure, @GPhys.
Hi @Krijn
I'm reviewing the entirety of my notes right now, but I never quite understood the inverse and implicit function theorem chapter (at least to my satisfaction)
I have a tangible question, but I'll finish what I'm working on right now first :)
I'm showing $f(x,y)=\sqrt{x^2+y^2}$ is not differentiable at the origin
I can explain the inverse and implicit function theorem to you if you want - or do you have a different question?
@GPhys Sure.
which I can prove doesn't exist since $$\frac{\partial f}{\partial x}=\lim_{h\rightarrow 0}\frac{\lvert h\rvert}{h}$$
which has limiting sequences of -1 and 1 which is enough to show it has no limit in general
(that's all just scraped from the definition of the partial)
using $$\frac{\partial f}{\partial x_j}=\lim_{h\rightarrow 0}\frac{f(x+he_j)-f(x)}{h}$$
user92578
f(x) = \int(0, 1)(\abs(x-t)dt
user92578
dammit how does this work?
11:45
That's alright, but you could also just directly derive it from the definition of a derivative.
Suppose we have the Maclaurin series of a function f, and it converges in a radius R. Then suppose we define a matrix argument to the function in a similar manner to the exponential definition of a matrix. Do all those functions converge respectively in the same radius R for all matrices?
user92578
Is there a syntax help page for the LaTeX in this chat?
@BalarkaSen ah
@BalarkaSen okay this would be great; let me grab the definitions my notes have for reference
I'll try making it a question in MSE
user92578
user92578
11:53
I have a function like that, where 0<=x<=2
@GPhys OK, sure.
user92578
I'm sorry for the non-LaTeX version
user92578
And I need to find the minimum and the maximum of that function
@BalarkaSen These are the two fundamental results of the chapter imgur.com/a/AOERN
user92578
So I would appreciate a small hint as to where I should start?
11:54
@GPhys Your link doesn't work.
@BalarkaSen does it work now?
oh, it says it should be, I'll just link the images directly
user92578
Would this perhaps work? Dividing the integral into two parts? math.stackexchange.com/a/1057449/272766
Not sure why you are linking those to me.
11:57
I'm just showing you what I'm looking at in the chapter so we're on the same page
OK, sure.
@Krijn Ah ok... Thanks a lot!! :-)
So @GPhys, let's start with the inverse function theorem. What's unclear about it?
@BalarkaSen I think I'm okay with the inverse function theorem. It seems straightforward
@robjohn I have found in my notes that $\widehat{sgn(x)}=2 PV \frac{1}{i \xi}$.
12:03
it's just establishing the existence of it?
Okay, so can you please tell me where you are having confusion?
@robjohn How do we show that the FT of $x$ is the derivative of the Dirac delta ?
@GPhys depending on the normalization of the Fourier Transform, that is probably true.
@GPhys Not sure what you mean. The inverse function theorem gives a sufficient condition for a local inverse of a function to exist, yes.
@Evinda what is the FT of the Dirac delta?
12:06
@BalarkaSen the questions of the chapter regard, say, computing the inverses; this provides a list of things I need to check to prove I could compute one
such as, let me start writing out one of them
@robjohn $\widehat{\delta}=1$
@BalarkaSen Okay I have $f:\mathbb{R}^2\rightarrow\mathbb{R}^2$ given by $f(x,y)=(x,y+h(x))$ where $h$ is a differentiable function of one variable. I want to compute $f'$, show $f'$ is invertible, then invert it. I can find $$f'=\begin{bmatrix}1 & 0 \\ h'(x) & 1\end{bmatrix}$$
From that we get that $\widehat{\delta' }=2 \pi i \xi$, right? @robjohn
what do I need to check to prove it is invertible? (I'm guessing its inverse is going to be the same thing with $-h'(x)$)
12:12
You need to check that determinant is nonzero.
Okay that is easy enough
It is $1$
You should prove why determinant nonzero implies invertibility though, if you don't know that.
Correct, @GPhys.
invertibility implies nonzero is a corollary of $\mathrm{det}(A)\mathrm{det}(B)=\mathrm{det}(AB)$
I am working on the other way
@GPhys Note that even after you do all of those, you still don't have a local inverse for $f$.
@BalarkaSen For exercise 2.1.9, in which we take $\Delta^n$ and identify all faces of the same dimension: Do we have that it has the homology of an $n$-sphere in odd dimensions and of a point in even dimensions?
12:19
@BalarkaSen okay I think I sketched an uglier proof of the other way
so it is invertible if and only if the determinant is nonzero
@AkivaWeinberger I don't think so. Take $\Delta^2$. If you do all the identifications, you end up with a contractible space.
@BalarkaSen how do I invert?
@GPhys You need to know that $h$ is continuously differentiable. That $f$ is $C^1$ is essential in the inverse function theorem.
@AkivaWeinberger Oh, I see you did say it has homology of a point in even dimensions.
Yes, that looks fine.
user92578
I need to prove that a semi-ugly function is the solution to the differential equation (y')^2-(y^2+4)(y'-1)=0. Is there something else that I can do other than just putting the function (and it's derivative) in and starting to simplify?
12:23
@GPhys Oh, you mean how to invert the matrix $f'$?
@BalarkaSen yes, inverting $f'$
Use the block matrix $[f' | I]$ and do row reduction.
so it is the same as $f'$ but with $-h'(x)$ instead
Yah.
Well, I should have known (can you guess why?).
Or maybe don't guess, it's trivial and distracts us from what we're trying to do: I should have known because it's an elementary matrix which adds something to the row. So if you want to get back to the previous matrix (so you multiply with inverse) you have to subtract that something.
So the inverse is $h'$ replaced by $-h'$. Anyway, move on.
@robjohn So we have to show that $\widehat{x}=2 \pi i \xi$, right?

But how can we calculate the FT of x? I wanted to use the definition, but the result tends to $+\infty$. So we have to use districutions. Could you explain to me how?
12:28
yes it is clearer to me now what and when I can borrow from my linear algebra
okay I still do not understand i.imgur.com/aQfqj8Y.png Specifically, how I'm supposed to be using the implicit function theorem to answer this
@GPhys Basically, you can think of inverse function theorem as follows.
but finish your thought first
Wait. $\Delta^2/\sim$ is contractible?
(This is the "dunce cap", right?)
Suppose $f : \Bbb R^n \to \Bbb R^n$ is a $C^1$ (continuously differentiable) map near $a$. Then IFT says if $f'(a)$ is invertible (this means $f$ is invertible at the infinitisimal level) then there are neighborhoods $U$ and $V$ of $a$ and $f(a)$ resp. so that $f : U \to V$ has a $C^1$ inverse.
@AkivaWeinberger Some people call it the dunce cap. Some people also call \Delta^2 with edges cyclically pasted dunce cap, which has nontrivial 1st homology hence not contractible.
But yes, this dude is contractible. Prove it.
Hatcher calls it the dunce cap, IIRC
12:31
@BalarkaSen okay this sounds good
@AkivaWeinberger Not sure if that was this one or the cyclic one.
@GPhys So the reason IFT is powerful is that it takes infinitisimal information ($f'$ and whatnot) and spits out local information (local inverse).
If you know some topology, then the statement of the theorem gets even more clear: if $f'$ is invertible, then $f$ is a local diffeomorphism (i.e., $C^1$ homeomorphism).
Unfortunately, the extent of my topology knowledge is limited to that appropriated to teach us metric spaces
You haven't encountered homeomorphism of metric spaces, then? Nevermind that, if so.
Metric spaces are nice. There's a metric space of (compact) metric spaces!
@AkivaWeinberger Yeah, Hatcher calls this dunce cap, not the cyclic one (unlike Munkres, e.g.)
12:35
Huh.
@BalarkaSen nope
Hm. Is this homotopic to a triangle?
OK, let's move on to your other questions.
@AkivaWeinberger Prove it.
I think it is. Start with the Hatcher-type dunce cap, and kind-of "unzip" the identified edges… I can visualize it, but I'm not sure I can describe it well or prove that it works
@BalarkaSen the book we used is open source, jirka.org/ra/realanal.pdf The beta of chapter 8 is being used to teach the current section of our class, and regards our current discussion
the last chapter of the book in the above link, 7, is what we did of metric spaces
12:38
There was this YouTube video of someone making a dunce cap with zippers. He had a big triangle of cloth, and zippers on the edges, and just zipped them to get the dunce cap.
but the thing at hand is imgur.com/aQfqj8Y when you're ready
@AkivaWeinberger You have the right idea. Here's the rigorous math: inflate the triangle into a disk. Then the thing is a cell complex, obtained by attaching a disk to a circle by the attaching word $x + x - x$ (you can figure out what that means). This attaching map is homotopic to the identity map, as $x + x - x = x$.
I think every stage of that zipping was homotopically equivalent to the starting triangle
Homotopic attaching maps give homotopy eq spaces (recall from chapter 0)
So you're through. Your space is upto homotopy eq a disk, which is contrct.
12:41
@GPhys You don't need to know implicit function theorem to do this :)
For 10a (taking a bunch of $\Delta^2\rm s$ and identifying pairs of edges, and proving it's locally homeomorphic to $\Bbb R^2$)…
@BalarkaSen :<
10 is subtle, @Akiva, be careful.
Can't I end up with a surface with boundary? Like, identifying no pairs of edges. That's not locally homeomorphic to $\Bbb R^2$ at the edges.
No there cannot be any open edge due to hypothesis.
12:44
Does it say I need to identify every edge with something?
Yes, that's what it means.
Arright.
Is this what's called a "manifold"?
I need to prove I get a 2-manifold?
@BalarkaSen I don't even understand what the answer desired is, or what I'm supposed to get out of it
@GPhys Suppose you have $x^2 + y^2 = 1$. How would you, naively, solve for $y$ in terms of $x$?
12:46
$\sqrt{1-x^2}$
So, possible problem points are edges and vertices, clearly. Edges are easy because they always get identified with another edge, so it's locally homeomorphic to a piece of a parallelogram.
or, $\pm$ that
(I.E. Two triangles stuck together)
where $+$ or $-$ works better for $(0,1)$ or $(0,-1)$ respectively
@AkivaWeinberger You need the identification information here!
@GPhys OK, now note that $\pm \sqrt{1 - x^2}$ is not a function.
So you need to choose.
Say I choose $y = \sqrt{1 - x^2}$ (positive sign).
That only makes sense for $x \in [1, -1]$, right?
There are only two ways to identify edges; preserving orientation or reversing it.
@BalarkaSen yes
Akiva: I mean you need the information that you're gluing pairs of edges here.
Otherwise you could just stick in three edges and you're out.
But it says that it's pairs of edges in the problem
12:49
Yes, I am emphasizing that it's needed.
Right. With pairs of edges, we get a little parallelogram thing, and there are no other simplices within a small neighborhood of any point on that edge, so…
…it's locally homeomorphic to $\Bbb R^2$ there.
@GPhys So $y = \sqrt{1 - x^2}$ does not work near say $(1, 0)$, because any nbhd of that point contains things $(a, b)$ with $a > 1$.
At the vertices, there are only finitely many simplices meeting at them (say, $n$), and there's a natural cyclical order to those simplices, so I can take a small circular neighborhood and map the part of that neighborhood in a given simplex to a sector in $\Bbb R^2$ of angle $2\pi/n$.
But it does work for $(0, 1)$, because we can choose a sufficiently small nbhd (namely, the open upper hemisphere) where points are $(a, b)$ with $-1 < a < 1$.
Wait, no.
I need to specifically point out that I can't get any points like you find on the double-cone.
12:54
Yes, @Akiva.
@BalarkaSen okay
But, like, you can't.
'Cause that would require the power to identify vertices rather than just edges.
Specifically, if we disconnect the two halves of that cone, there are the same edge identifications.
Yes, things like that don't happen in spaces of question.
So there can't be any double-cone points.
Not the sort of proof that you'd program into Coq, but I think it's rigorous enough
@GPhys But does $y = \sqrt{1 - x^2}$ for $(0, -1)$?
12:59
@BalarkaSen no, we need $y=-\sqrt{1-x^2}$ for $y<0$
@Akiva To be entirely rigorous you need to prove that the link of a vertex in the delta complex of question is a cyclic graph.
But that's fine.
@GPhys Right. So what more is there is to it?
You have just solved the problem.
This is a (standard) example where the implicit function theorem holds (i.e., we have written the circle locally as graphs of functions, even though it is not a graph globally - it's a level set). But you don't need IFT to prove it.
The implicit function in general says for a sufficiently nice function $f : \Bbb R^n \to \Bbb R^m$ and for a sufficiently nice point $a \in \Bbb R^m$, the level set $f^{-1}(a)$ is locally graph of some function over the $(n-m)$ plane.
(punchline for when you learn more topology: this says sufficiently nice level sets are manifolds)
@Evinda No... $\widehat{x}$ is the derivative of the Dirac delta
@Evinda distributions are tested against test functions
okay
What's next?
that's the last section of the stuff on the exam (or that we've covered), let me work back to the last previous thing that needs....work
I can just note that $1,t,t^2,\ldots ,t^n$ are a basis and then count $n+1$ items?
13:15
Yes. You also need to prove they are a basis, but you can see that I guess.
it's a basis if their linear combination is the span, which is pretty much directly by construction
so yeah
okay
No, you need to show they are linearly independent. That's what I meant.
It's clear, but make sure you don't skip this.
oh
A basis is a thing which spans the whole space and it's elements are independent.
what I suggested would only show the dimension is $\leq n+1$
13:18
Yes.
they are linear independent because
if they weren't then I could represent one of them as a linear combination of the others, which is impossible
Because $t$ is a variable, i.e., a transcendental and cannot satisfy a polynomial equation. But yes.
e.g. $t^2$ can't be written as a linear combination of $1,t,t^3,t^4,\ldots$
okay
I don't have a grasp for the operator norm
$\max_{\|x\| = 1} T(x)$?
13:24
Ok, sup. Close enough.
Well, they are the same.
'Cause $\|x\| = 1$ is a sphere, so compact, and by max value theorem, $T$ attains a maximum.
I understand what $L(X,Y)$ is, on the definition level. $A$ is in $L(X,Y)$ if $A: X\rightarrow Y$ such that if $a\in\mathbb{R}$, then $A(ax)=aA(x)$ and $A(x+y)=A(x)+A(y)$
(They are the same when $T$ is a linear transform on $\Bbb R^n$: I just noted we are working with general normed spaces (are we?))
yeah this is not specific to $\mathbb{R}^n$
What do you not grasp about this norm?
I remember doing this question for homework a while ago
and it took me a very long time
and, I guess
I don't understand what is motivating the operator norm, as defined here
13:34
@GPhys Well, it turns the vector space $L(X, Y)$ into a normed space, for starters.
so, reworking it, I think the answer to a) is $2$
I'm not sure I get what 10b is asking
You're correct. Can you compute norm of a diagonal matrix with diagonal entries $d_1, \cdots, d_n$ in general?
"Show the edges can always be oriented so as to define a ∆ complex structure on the quotient surface. [This is more difficult.]"
@AkivaWeinberger Note that in the definition you just pasted a bunch of simplices, not carding about the orientation on the edges.
So why is this thing a $\Delta$-complex? Can you orient the edges so that it becomes one?
13:37
@BalarkaSen with a euclidean norm on $\mathbb{R}^n$ ?
Yes, @GPhys.
with some Lagrangian multipliers I think eventually I can get there
the $\mathbb{R}^2$ case is easy
You don't need Lagrange multipliers. Do it simply.
How did you do it with (a)?
Why does the orientation matter? Am I missing something obvious?
@Akiva In $\Delta$-complexes, you glue simplices preserving orientation.
A simplex with edges pasted cyclically is not a Delta-complex as described.
13:40
Ah…
So that's what that paragraph on page 104 (where he discusses the cyclic dunce cap) was about
I completely missed that
Yes, and that's why you shouldn't skip. But note that doing exercises actually point out which things you left out!
@BalarkaSen if I call the matrix $A$, and let $x=(a,b)$, $Ax=(a,2b)$, then the norm of that is $\sqrt{a^2+(2b)^2}$, but $\sqrt{a^2+b^2}=1$ by restriction, so I can write $\sqrt{1-b^2+4b^2}=\sqrt{3b^2+1}$ which has a supremum of $2$ over the restricted range of $b\in[0,1]$. That is, $b=1$ is the max by observation
So the conclusion is that trying to learn the theory by doing the exercises is a much more efficient technique of learning than just learning the theory, which can be dry and unmotivated and frustrating.
am I doing this the hard way?
I didn't skip that section. I just didn't understand it when I read it and then forgot it
13:42
Well, now you do. My point stands.
@GPhys Yes, way too hard. Note that $\sqrt{a^2 + (2b)^2} \leq \sqrt{(2a)^2 + (2b)^2} = 2$. Now note that $2$ is actually attained: take $v = [0, 1]$. $Av = [0, 2]$, which has norm $2$.
So $\|A\| = 2$.
The most common gaps between primes are multiples of 3 according to Denis Ivanov.
@BalarkaSen oh, $\max \{d_1,d_2,\ldots ,d_n\}$?
13:48
I did it the hardest way
subspace of a vector space is not always a vector space?
oh
it needs $0$ and additive inverses aaaand that's it?
I'm back (briefly). Assign a random number to each vertex and let the edges point from smaller to larger
if a subspace has $0$ and an additive inverse for each element, then it is also a vector space?
@AkivaWeinberger Works.
@GPhys $\{1, 0, -1\}$ in $\Bbb R$ has $0$ and an additive inverse for each element in it.
Is it a vector subspace?

« first day (2063 days earlier)      last day (3254 days later) »