« first day (2298 days earlier)      last day (3021 days later) » 

13:33
If anyone get then tell me(or any link where can I get it's proof)
@Ramanujan use induction and the formula $\sin(2\theta)=2\sin\theta\cos\theta$
@DHMO @DHMO yes, i meant f(f)
alright
@DHMO any proof? Iam not trying check,trying to proof
eh
13:37
@Ramanujan the proof is induction by the formula $\sin(2\theta)=2\sin\theta\cos\theta$
trying to proof doesn't include getting it spoonfed. altho it might help to read similar proofs. (to get the identities straight)
as DHMO said, induct your way through it
@DHMO wait a minute
23rd one
@Ramanujan use the identity you just proved
Before seeing hint I don't know that identity
actually the identity is a little bit different
13:42
then multiply $\sin\left(\dfrac{2\pi}{2^{64}-1}\right)$ to the whole expression
powers of 2, not simply n
$2^n$
@Null ?
it is $2^n$ in both instances
i dont see here powers of 2
@Null I see clearly $\cos(2^n\theta)$
oh well
13:44
The notation is poor, but it's the same problem
yes
@teadawg1337 which notation is poor?
@DHMO do you mean multiply and divide or multiply on both sides?
@Ramanujan whichever you like
@DHMO The lack of parentheses in the original image
13:45
@teadawg1337 I see
cfp
cfp
@Semiclassical Just noticed the bounty you placed on my question. Thanks again, it's much appreciated.
@DHMO and then multiply and divide by 2?
@Ramanujan sure
Oh dear, I just noticed that the textbook being used has the same notation issues
How can the Cauchy-Schwartz inequality be used on sequences of numbers that might not have "enough" elements?
13:49
@Null for example?
@DHMO iam left with this…then?
@Ramanujan it should be $2^{64}$ not $2$
@DHMO in denominator?
yes
for $a_i$ we take the numbers 1 to 10. For $b_j$ the numbers 20 to 50. Would that be simply not intended? @DHMO
(and n= higher then 10)
13:51
@Null then the inequality does not apply?
you can also fill the other sequence with zeroes
where did the denominator of the first line come from?
I will show full Sol
@DHMO this iam getting
@DHMO are you there?
@Ramanujan I am not getting the first step
Hello omeone know the Lusternick-Schnirelman theory ?
14:07
@DHMO for each angle of cos I multiplied and divided with sin of that angle ,and then to get in 2sinacosa form I multiplied by 2^(no.of terms)
@Ramanujan fair enough
now you can continue by using $\sin(\pi-x)=\sin x$
In mathematics, the Lusternik–Schnirelmann theorem, aka Lusternik–Schnirelmann–Borsuk theorem or LSB theorem, says as follows. If the sphere Sn is covered by n + 1 open sets, then one of these sets contains a pair (x, −x) of antipodal points. It is named after Lazar Lyusternik and Lev Schnirelmann, who published it in 1930. == Equivalent results == There are several fixed-point theorems which come in three equivalent variants: an algebraic topology variant, a combinatorial variant and a set-covering variant. Each variant can be proved separately using totally different arguments, but each variant...
@DHMO do I need to know the value of 2^64 - 1?
@Ramanujan no
Then how?
1 min ago, by DHMO
now you can continue by using $\sin(\pi-x)=\sin x$
14:10
@DHMO i want to know if we can apply this theory on ordinary differential equation ?
@Vrouvrou no idea
@DHMO I can't find any term of sin(π-x)
@Vrouvrou How do you plan to apply this to ODE's?
@Ramanujan use $\sin x = \sin (\pi-x)$
@DHMO the LS theory is not quite limited to that theorem
14:12
@BalarkaSen I'll shut up then
@BalarkaSen i read many papers about this theory and all apply it on pde , i want to know if it can be apply it on ode
I have never heard it being applied to PDE's. By LS-theory I presume you meant the study of LS-category.
@DHMO NB I don't know much about it either
@BalarkaSen yes they obtain cat(\Omega) of solutions
@DHMO it's not getting reduced
@Ramanujan apply it only to one of the sines
no
use $\sin x = - \sin(2\pi-x)$
14:17
Instead of using that identity, I believe it would be easier to observe the period of $\sin(\theta)$. The inclusion of $\pi$ in the problem isn't a coincidence
@BalarkaSen Heyo
many people here
Hi @Danu.
Still complex geometrizing hard?
Yeah man
I'm almost done!!!
9 pages 'til I finish chapter 5 of Huybrechts
All the big theorems rolling in now
it's great :D
14:19
All the $H^q(\Bbb C\rm P^n,\mathcal O(k))$, done :D
This Kodaira vanishing theorem is excellent :D
Numerator
how is this any different from cryptology haha
@Danu Yeah, I vaguely remember it
@Ramanujan nice
@BalarkaSen Now I'm doing that "Weak Lefschetz Theorem" you mentioned a while back.
14:21
Then?
Ah I see
(relating cohomology of a hypersurface to that of total space)
@Ramanujan the you can cancel
What proof did you see?
This is all SES's of sheaves
I know there's a Morse theory proof, and if I keep reading Milnor I'll see it.
14:22
oh, really?
He talks about complex geometry?
I really wanna know about Morse theory.
@DHMO really thanks for help,i got now,so I think we can prove that formula in the same way,again thanks
it's cool stuff; i started reading some for passing time
@Ramanujan you are welcome
14:24
@BalarkaSen please do you know what is the LS category of an interval ?
@Vrouvrou you should be able to answer that yourself if you know the defn
@DHMO btw how you got that we need to use sinx=-sin(2π-x)?
@Ramanujan because $2 \times 2^{64} - 2 = 2(2^{64}-1)$
@BalarkaSen i think tht is 1
it's 0. the ls-category is 1 less than the number of open sets you cover by
14:31
how is a piecewise function called that only fills in the gaps? like for (sin(x)/x), y=1 for x=0?
(point gaps i mean)
So in physics class we were discussing the lift force, $F_L$, and specifically its $y$-component $F_{Ly}$
I think it's great that the force that makes planes fly is called $F_{Ly}$
@AkivaWeinberger no, you are kidding now? :s
64/16 just cancel 6 you are left with 4/1
Hopefully $F_{Ly}=-mg$ or you're gonna have a bad time
(or you're landing)
14:35
@Ramanujan Woah hold on, where are you getting 64/16?
@DHMO why in particular is sin(36) good to know?
@teadawg1337 His brain?
@Null because it's nice
@Ramanujan That's also not how you simplify fractions...
@Akiva "$m$y $g$od [and more screams]"
14:36
@teadawg1337 I read that somewhere
@teadawg1337 whoosh
@DHMO well -1 roughly will now be in my head, congratz on that :P
I thought it was in reference to the problem that was being worked on previously
@teadawg1337 It's a sort of mathematical joke — $\dfrac{\cancel64}{1\cancel6}=\dfrac41$ is the wrong way to simplify fractions but gives you the right answer anyway
(LaTeX help? How to do diagonal strikethroughs?)
@teadawg1337 hah,that problem was solved(over)
14:38
@BalarkaSen i have this definition : If $Y$ is closed subset of a topological space X , the LS category is the least number of closed and contractible sets in X which cover Y
I'm aware of the joke, I just didn't realize it was a standalone statement at first
TeX: {
extensions: ["cancel.js"]
} $\cancel6$
Aw, didn't work
How most people research maths:
1. Encountered an or a bunch of interesting problem
2. identify a natural structure in the problem
3. Describe the structure with axioms
4. Prove and derive theorems
5. When the framework is established, generalise or abstract to similar objects
6. Rinse and repeat

How I research maths:
1. Choose a structure or mathematical object
2. Generalise it by discarding axioms
3. Use this new set of axioms to construct a new object
4. Make sure the constructed object exists and nontrivial
@Vrouvrou Ok. That's not the standard definition, which is 1 less than the number of open/closed sets.
@Secret I prefer the bottom way lol
14:42
then if i use this definition the LS category of an interval is 1 @BalarkaSen
Both approaches are important. The top approach is important when solving real world problems. The bottom approach is more like pure maths/fine arts approach and focus on building as many of these structures and then describe them later.
The bottom approach is more experimental oriented, possibly because of my physics background
Most of the time when I researching on things, I am usualyl drawn to how instrinsically interesting they are. To me, applications are often just an excuse of making them more tangible so I can "mess with it with my hands"
@BalarkaSen thank you
15:09
@Secret so you are lego, but others playmobile?
possibly, although I can switch between the two modes depending on the problem in question
Hi chat
Good weekend all users! Bye.
My argument appears to be invalid. Hopefully I can fix it later.
@Astyx hi
15:29
How would one prove that if $f:\Bbb R_{+}\to \Bbb R$ satisfies $$f(nx)\to_{n\to +\infty}0$$ for all $x\ge 0$, then $$f\to_{x\to +\infty}0$$
@Secret That is not how most people research math. Certainly the word 'axiom' doesn't show up in anything I've ever written, or am liable to.
In that case there is a lot more to learn cause I am only familiar with one of the ways to study abstract algebra (possibly how I made that mistaken claim above).
I do not study abstract algebra :)
How do researchers study topology?
(Most people in algebra study things like algebraic geometry or commutative algebra, where a lot of the time the interest isn't in the structure of some axiomatized object, but some more geometric question)
15:36
i see
@MikeMiller This is an example of a self-contradictory statement
"any papers" :P
I was imagining if Akiva would say that.
@Secret Sorry for not responding, I intend to. But I'm a little busy this morning.
it's ok, take your time. I am also in the middle of something
15:42
@BalarkaSen Say what?
Ohh
He didn't say 'axiom', he said ''axiom'', doesn't count
@Secret So I think it's impossible to say something about how "topologists" work. I can say something about how I work. I have questions that I think are interesting (which 3-manifolds embed in $\Bbb R^4$? is one of my favorites!), and there is a broad spectrum of tools which have historically been developed. So a lot of my thinking is either towards extending those tools to be more useful, or finding ways to use them. Sometimes I have a flash of insight and think
"Ah, I can use (idea) to try to tackle (question)", and think about that for a while, and then it doesn't work.
2
Hmm, it sounds kinda similar to organic synthesis in chemistry, where the interesting thing is a target molecule to be synthesized. There are theories and reaction pathways historically developed, and our job is to optimise these pathways in order to make the molecule more efficiently
Therefore to me, it seems like the maths tools will be used to figure out which 3 manifolds will embed in $\mathbb{R}^4$ or to better characterise them
15:59
Has one of you studied in France ?
And were one really genius they could stumble upon an altogether new tool
16:11
Is any path-connected space $X$ with $\pi_n(X) = 0$ for $n > 1$ homotopy equivalent to a CW-complex?
Yes.
Actually, um.
I actually don't think it is.
Look at say the Warsaw circle. I believe that has higher homotopy groups all $0$ just fine.
weakly contractible does not imply contractible
how is that relevant?
16:23
what happens if a weakly contractible space is homotopy equivalent to a CW complex?
oh, I thought that was a reply to my example. yes, it's contractible.
Any hints? f(t) is continuous and t^-2/3 is continuous on (0,1] so if we take the limit from the right the product should be continuous on [a,1] but I am not sure how to formalise this
@BalarkaSen Thank you. This is a good counterexample. :)
Btw. when reading about mostow's rigidity theorem, nobody mentions the connectedness. Is it usual for people in this subject to only consider connected manifolds?
So mostow's rigidity theorem only holds for connected manifolds?
16:33
but a proper definition of hyperbolic is "has universal cover $\Bbb H^n$", anyway
oh ok
i mean, for disconnected manifolds you just need to check componentwise
isnt it a riemannian manifolds with constant sec. curvature -1?
Oh okay. I see. So it does hold for disconnected manifolds :)
when $M$ is connected, the two statements are equivalent
there's just never any interesting reason to consider disconnected things when one can always say "check components"
makes sense, thank you
16:37
@GridleyQuayle Since $f$ is continuous on $[0, 1]$, you can bound it by a constant (why?). Work with that.
@BalarkaSen wow, nice
f is bounded, by M say, by the Extreme value theorem. hence $ t^{-2/3} f(t) \leq Mt^{-2/3} $ Integrating we see that the RHS converges to 3M (finite) thus the LHS converges.
Thanks @BalarkaSen
You have proved the limit does not blow up to infinity. There's still an issue about showing it actually exists.
I suppose it's theoretically possible that the integral has a weird oscillation thing as $a\to0$
@BalarkaSen let the boundary be $[m,n]$.
$m \le f(t) \le n$
$mt^{-2/3} \le t^{-2/3}f(t) \le t^{-2/3}n$
$\int_a^1 mt^{-2/3}\ \mathrm dt \le \int_a^1 t^{-2/3}f(t)\ \mathrm dt \le \int_a^1 t^{-2/3}n\ \mathrm dt$
$3m \le \int_a^1 t^{-2/3}f(t)\ \mathrm dt \le 3n$
and there is where i got stuck
16:43
like $\sin(1/x)$
same place as @GridleyQuayle
Show that it's a Cauchy sequence or something, probably. For a given choice of a sequence $a_i$ tending to $0$.
Should be doable.
Maybe look at the derivative wrt $a$?
I don't know if it works, but it's an easy computation, right?
@AkivaWeinberger but it is not continuous at $x=0$
@DHMO He's speaking that it's theoretically possible until you prove it.
16:45
@DHMO It's not defined there, that's my point
@BalarkaSen I know
@AkivaWeinberger but f is continuous on [0,1]
@DHMO But $\int_a^1t^{-2/3}f(t)dt$ isn't defined at $a=0$
So theoretically that function could look like $\sin(1/a)$
@AkivaWeinberger i'm saying that the question requires $f$ to be continuous on $[0,1]$
I think it's easy enough to show that it's a Cauchy sequence, @Gridley.
It's literally the same argument.
That what is Cauchy? You mean consider a sequence of integrals?
16:48
@DHMO So?
$f$ isn't the function I'm worried about
@AkivaWeinberger so $\sin(1/x)$ doesn't satisfy that requirement
oh
@DHMO I wasn't comparing $f$ to that function, I was comparing the integral to it
So $\lim\limits_{x\to0} \sin(1/x) x^{2/3} = 0$
Let $E = C^0([0,1], /Bbb R)$ with the infinite norm $||\cdot||_{\infty}$ and $\phi:\begin{cases}E\to \Bbb R\\ f\mapsto \int_{0}^{1/2}f - \int_{1/2}^1 f\end{cases}$, and finally $H = \ker \phi$. How should I go about computing the distance from $id$ to $H$ ?
@Gridly Right.
16:50
We can define a function based on $x^{2/3}\sin\left(\dfrac1x\right)$
would it be a counter-example?
no, because $\displaystyle\int_0^1 \sin\left(\dfrac1x\right)\ \mathrm dx = 0.504\cdots$
Thanks, I'll give that a shot. Another quick question, for a space to be Banach, must every cauchy sequence converge uniformly, or just converge?
This thing (the integral) has a bounded derivative, right?
The derivative being $-a^{2/3}f(a)$.
just converge. "converge uniformly" doesn't make sense in an arbitrary metric space
@AkivaWeinberger bounded derivative implies convergent integral?
Maybe?
Intuitively it seems true
16:53
I do not rely on intuition
As well you shouldn't
isn't it enough to show that the improper integral is absolutely convergent? it should imply the convergence of the improper integral itself
and the argument gridley gave earlier shows that it's absolutely convergent, no?
Sure, that's what we are doing.
@prodprod No? It just says the absolute integral is bounded.
well ok, technically you still have to say a bit more
namely that the integrals of the absolute values increase as $a$ approaches 0...
17:11
is that roughly correct?
So, MathCamp is asking us if we have any original math puzzles we can send them for them to put on this year's Qualifying Quiz. So I now might want to solve the meta-problem of coming up with an interesting problem.
This raises the question — what other meta-problems are there?
What kind of puzzles ?
ask for a proof of uniformization
@Astyx For reference, these are the puzzles from last year: mathcamp.org/2016/qquiz.php
I've solved the 2014 ones
@GridleyQuayle I didn't look at that but you seem to have proved that $I_{1/n}$ is Cauchy. What I want to do however is to show is $I_{x_n}$ is Cauchy for any sequence $x_n \to 0$. The thing is you want to show $I_a$, as a function of $a$, is Cauchy-continuous in $[0, 1]$, which might be easier to do. Then continuity is a corollary.
I'll warn that I haven't tried this approach, so no guarantees it'd work.
k, gotta go
17:43
Hi @MikeMiller
I'm trying to get better at understanding SES's of vector bundles.
you put the one thing in then you get rid of it
easy as that
haha
Sure.
Given a SES $0\to L\to E \to F\to 0$ where $L$ is a line bundle, I'm trying to understand how to get $0 \to L\otimes \bigwedge^{p-1}F\to \bigwedge^p E\to \bigwedge^p F\to 0$
I thought I should do it through understanding transition functions, maybe.
Then I realized I didn't understand transition functions of $\bigwedge^k E$ in terms of those on $E$.
start topologically. if $E = F \oplus L$, we have $\Lambda^p E = \Lambda^p(F \oplus L) = \oplus_n \Lambda^nF \otimes \Lambda^{p-n} L = \Lambda^p F \oplus \Lambda^{p-1} F \otimes L$
so hey, it works out
that's all i got for ya
17:50
Hmm...kay
be my guest on transition functions
Okay guys. I'm trying to come up with the best way to get feedback on my Undergrad study activities and mock exam questions. Do you think I could post my workings to SO just to get feedback (and obviously corrections if needed)?
I dunno, try it, see what happens
Put an "I'm not sure if this belongs here" disclaimer on it
I'll give it a go!
Good idea!
I would vote to close immediately. Try reddit.
18:02
Mike! You're a legend, was trying to work out where to put it!
@MikeMiller My thinking was, even if it got voted to close, people would comment saying where it should get asked instead.
18:48
@Ramanujan I was just curious about the problem you were discussing..
19:03
@Danu Out of curiousity: Of the complex geometry stuff you're doing (Huybrechts etc) is any of it in the realm of Hodge theory?
@Semiclassical Hodge theory is essentially a black box in this book.
19:23
Do one-to-one function and one-one function both mean an injection?
I hate it when people use that terminology :P
ha, which one?
@Brody both are terrible
just like onto
I agree
(not that anyone cares)
@Tobias I can see how one(-to)-one function is annoyingly close to one-to-one correspondence but I don't find onto too offensive
19:32
@Brody Consider the statement "$f$ is a function from $X$ onto $Y$" and note how easy the word is to miss
@Tobias Ohhh, definitely
19:46
@Semiclassical note that "Hodge theory" and the Hodge whiten are different
the Hodge theorem*
@MikeMiller I read that as the Hodge within
20:08
definitely much cooler
20:20
@MikeMiller Which should I be thinking of if I'm seeing stuff about Hodge filtration? (The first?)
@Brody Yep.
@Semiclassical Yes. Danu thought you were talking about the latter.
@MikeMiller @BalarkaSen I have a question for you
If $X$ is a continuum say that $x r y$ if $x$ and $y$ are contained in a nowhere dense continuum of $X$.
$r$ is an equivalence relation.
Can you think of a metric continuum with a finite number $n>1$ of equivalence classes?
in $[0,1]$ each eq class is a single pt
but I want finite number
greater than 1 because $[0,1]^2$ is an example with only 1
it is a tough problem I think
fun to think about
20:36
Look at, suppose, the topologist's sine curve $\times$ [0, 1]. How do you plan to "connect" stuff in the accumulation plane of the sine wave to any point on the sine wave with a nowhere dense continuum?
you can do it
Oh, just join by a topologist's sine curve inside it
Very cute problem.
yes it may even be unsolved
there is a more difficult variation that I know is unsolved
so I wanted to look at this simpler problem
20:39
Neat. You want to study connectedness by joining with nowhere dense subcontinua instead of by paths.
You can also ask for joining by maps from continua with nowhere dense image in general, like you do for paths, say.
in an indecomposable continuum the equivalence classes are just the composants
like buckethandle
so its really only interesting for decomposable
anyone know how to prove $\displaystyle\sum_{k=0}^n {2k \choose k}{2n-2k \choose n-k}=4^n$?
20:49
more symmetrical: $\sum_{k+r=n}\binom{2k}{k}\binom{2r}{r}=4^n$
@arctictern this looks like induction. Thanks
that would be my approach
good afternoon
21:03
I have a proof of this, but it is a bit ad hoc. Consider the function $f(x)=\sum_{n=0}^\infty {2n \choose n}x^n=(1-4x)^{-1/2}$ then you get that identity by considering the Cauchy product $f(x)^2$
right
that was the first thing that came to mind, was too lazy to look up generating function f(x) though :P
Surely there's some long and complicated combinatorial proof.
21:38
so a transistor is the fundamental part of your computer that holds a 0 "no current" or a 1 "current" in the cpu and memory of all computers.
my question is does a tertiary transistor
one capable of 3 independent states
must be larger then a binary transistor?
this question stems from the end of moores law and information theory which shows tertiary computing is faster then binary computing
i.e hypothetically in the optimal for both cases does tertiary computing really allow for a greater upper bound in processing power of a cpu of certain size then binary transistors.

« first day (2298 days earlier)      last day (3021 days later) »