« first day (2777 days earlier)      last day (2261 days later) » 

12:00 AM
Everyone seems to have different answers to this stuff so it's not that obvious
despite the tone of your response
 
Your question is so vague and undefined that the only possible answer is to be flippant; if you have an actual question, you should ask that question.
 
trying to understand the differences in terminology between things like sentence, statement, expression, atomic, predicate, proposition, formula, etc
"free variables", quantifiers and what they are actually modifying, what these modifications are called, etc
I want to get my terms straight so I can speak about these things more precisely
But online resources are woefully inconsistent
And even when I read something like Tao's Analysis he doesn't get into these details too much
 
 
2 hours later…
1:46 AM
@user539262 Obviously an analysis text wouldn't go into this
You'd need a text on mathematical logic
or model theory or whatever
I don't think any logicians frequent the chat, but there are several logicians active on the main site (Mathematics)
 
@AkivaWeinberger um, @AlessandroCodenotti ...
 
he does set theory
 
Yeah, but that's not the same thing
though there is a big overlap, I guess
 
they're probably close enough for whatever that guy wants
@BalarkaSen you really like that one page from your dynamics book
 
Zee
2:44 AM
How do I show there is a canonical isomorphism between V tensor R1 to V
 
@Zee what have you tried
 
Zee
Well a bunch of stuff , am suspicious the question is wrong
 
it's not
 
Zee
Well one thing I was trying to do is show that V X R1 is isomorphic to the linear functionala from V to R1 but that can’t be right
 
hint, a \otimes v "=" a.v, where . is scalar mult
or v \otimes a, dunno why you wrote it backwards
 
Zee
2:47 AM
Sorry , can you write this in English? I can’t compile latex
 
You don't need to compile tex
 
Zee
Your not being helpful ...
 
ok
the solution is there, so google what \otimes means
 
Zee
Otimes is the tensor product ...
 
so what's the issue
 
Zee
2:56 AM
The correspondence you gave me is not a bijection
maybe my question wasn’t clear , let me restate it : V tensor R1 is ismomorphic to V , where V is a finite dimensional vector space and R1 is the real line
 
I didn't give you the full solution
 
Hello. How can i construct a linear transformation which maps real axis to a circle? I think its a linear fractional transformation, but it was not mentioned explicitly in the text.
 
@Shobhit this should be in any complex analysis book
 
Zee
The hint you have is useless , am not here to get spoon fed
 
I'm useless :'(
 
Zee
3:05 AM
I didn’t say that , your hint is though
 
why is it useless
 
Zee
It’s too trivial
 
lol
this is a trivial problem
 
If it's trivial for you, then what you're trying to prove should also be trivial to you
 
Zee
I don’t think you know the answer
 
3:06 AM
@0celo7 i am reading a complex analysis text, namely, complex analysis by lars ahlfors, but no hint is given as to how to achieve it. The author briefly mentioned cross-ratio to the reader, i think we are supposed to use that, but i dont see how.
 
@Zee maybe, maybe not
but you're not being very nice
 
The hint of @0celo7 is on point. He said that v tensor a "=" av, what if you take that "=" to try and define an isomorphism?
 
Zee
Am not being nice becouse I have 2 midterms in a week and three HWs to do and I don’t have time to listen to your half backed hints
 
I thought you didn't want to be spoonfed
I can give you the solution if you need it
 
@Zee you're the one who's trying to get help
 
Zee
3:08 AM
Yes and I regret it , wasted half an hour
 
(btw google finds the full solution immediately)
 
@0celo7 help please
 
@Shobhit Sorry not in a complex mood. Have you tried another book like Stein or Conway?
 
@0celo7 :( ok. I'll look into other books.
 
Zee
Can you post the link , I can’t find anything on google
 
3:12 AM
Say the magic word
 
@Shobhit consider the "Cayley map", $z \mapsto \displaystyle \frac{z-i}{z+i}$
 
Zee
please your highness
do you define the isomorphism by multiplying basis vectors by scalers ?
 
@Zee yes! that's what I said by the way (and what @0celo7 was hinting at)
 
@MatheinBoulomenos ok, i'll google that and get back to you :)
 
Take the map R x V --> V given by (a,v) --> av and lift to the tensor product via the universal property. This is an isomorphism.
 
Zee
3:14 AM
YA I thought about that
but I got stuck
 
You just have to prove it is one. It suffices to find an inverse.
 
Zee
Couse I didn’t think of using basis vectors to show bijective
 
You shouldn't have to use a basis
 
you don't need to involve a basis at all
(sniped)
 
Zee
How else are you gonna show bijectivity ?
 
3:16 AM
In this case you can write down the inverse map directly.
 
Zee
Ya , this seems too easy
I hate when this happens
 
you can also show that it is injective and surjective
 
Zee
I can’t believe I waste a whole day on this , especially since I considered this hint at the beginning and convinced myself it can’t be true
Anyway thanks 0celo7 and mathein
 
@MatheinBoulomenos on the wiki page of cayley transform, its written " mobius transformation permute the generalised circle in the complex plane", what does it mean? And i get how they made the transformation using given points, but i dont see how it maps the real axis to a circle. Help please.
 
generalized circles are circles and lines
 
3:26 AM
It means generalized circles map to generalized circles, and the same is true of the inverse map
 
but you don't need to think about how Möbius tranforms act on generalized circles in general for this problem, this is more concrete. If $z$ is real, what is the absolute value of $\displaystyle \frac{z-i}{z+i}$?
 
$z$ gives the real line, $iz$ gives the imaginary axis, $iz+1$ gives the line of points with real part 1 (as $z$ varies through the real numbers)
 
I found it funny that every time someone in the math chat get converted from field A to field B, a corresponding starred message popped up
6 hours ago, by Alessandro Codenotti
Was Mathei transformed into an analyst? What kind of sorcery is this!
 
@MatheinBoulomenos i get 1. So since mobius transformation maps generalised circles to generalised circles, the real axis will be mapped to a circle or a line, but { 1,-i,i} do not lie on a line, so the real axis get mapped to a circle passing through these points. Is this correct? @AkivaWeinberger
 
Make me wonder, can we do a category theory of people swapping between maths fields
 
3:33 AM
If we invert that line (sending it through the mapping $z\mapsto1/\bar z$) we get $\frac1{1-zi}$, which would be a circle
The one centered on 1/2 with radius 1/2
 
@Secret I'm not converted to an analyst. I just answered a question about analysis and Alessandro made a joke about it
 
Oooooooooo
 
Doubling it and subtracting $1$ gives us $\frac{1+zi}{1-zi}=\frac{z-i}{z+i}$, which gives the unit circle
 
Anyway, a category theory of people will probably be so complicated that nobody can verify the proofs
 
@Shobhit Yeah, that's the way Mathein was going for
(Those are the values at $z=\infty,-1,1$, I guess)
 
3:37 AM
@AkivaWeinberger got your analysis of the transformation, it was nice.
@AkivaWeinberger @MatheinBoulomenos thank you.
Yes the values are correct @AkivaWeinberger
 
Zee
@MatheinBoulomenos @0celo7 how can that be a bijection ? Take (1) (1 0) and (1/2) (2 0) these give you the same vector
 
yes, but they're also equal in the tensor product
 
Zee
Couse the tensor product is modded by pulling out scalers ?
 
@Zee Remember that ax⊗y=x⊗ay
 
@orbit-stabilizer Thank you so much for this reply.
 
3:49 AM
=a(x⊗y)
 
Zee
I see
 
@Silent, no problem. Was this a question on your analysis homework?
 
Zee
I don’t like like the book Lee smooth manifolds , it’s terrible honostly
 
You're proving V⊗R=V for vector spaces V, right?
You should know that V⊗(R^n)=V^n as well.
 
Zee
Ya
 
3:51 AM
When I upvoted an answer in main site, a message in a blue strip was displayed, that went away so quickly, i could not read. some initial words were 'you haven't voted in long time ' or something. Any idea what does that say?
 
And (R^n)⊗(R^m) is isomorphic to R^(nm).
 
Zee
This one I knew but the one before it seems interesting
 
Take n=2
 
@orbit-stabilizer No, no. Rudin's problem 4.8 says that 'f uniformly continuous function on real bounded set, then image is bounded.'
So, i had that question.
 
(v,w), an element of V^2, can be mapped to v⊗(1,0) + w⊗(0,1)
where (1,0) and (0,1) are elements of R^2
(Note that it's not a "pure tensor")
 
3:54 AM
If f is a continuous function defined on a compact set, is the image bounded?
I'd like to say yes
We have uniform continuity due to the compactness
 
Zee
If F is real valued then yes
Continoues function takes compact to compact and by hein Borel compact is bounded
 
(In the space V⊗W, things of the form v⊗w with v∈V and w∈W are called pure tensors. Things of the form v1⊗w1+v2⊗w2 are not necessarily equal to pure tensors, but they're still in V⊗W.)
 
Zee
Thanks akiva , tensors are new territory
 
(Those are meant to be subscripts.)
 
Zee
Are we talking about tensors as an algebraic structure or as something more specific ?
 
3:57 AM
@orbit-stabilizer Yes, it's part of the reason compact sets are "almost finite" (not a rigorous term)
@Zee Tensor products of vector spaces
 
Okay. But, if we know that for f, the preimage of every open set is open. So find some open cover of Im(f), look at preimage of that, that gives us an open cover for the domain of f. By its compactness, we have a finite subcover. Now map that back, and we get a finite subcover of Im(f)
Does this work?
 
Okay. So that gives us compactness. And apparently in some topologies, compactness -> boundedness?
I don't know much topology
Can we even talk about boundedness without a metric?
 
In Euclidean space $\Bbb R^n$, something is compact iff it is closed and bounded.
 
Yup, by hiene-borel
 
4:00 AM
In a general metric space, I think something is compact iff it is complete, closed, and totally bounded?
 
But, I'm interested in maps of the form f:X->Y where X and Y are topological spaces
 
Zee
There is a Version for metric spaces but it has weaker conclusion
Yes akiva
 
"Boundedness" on its own is not a topological property
so you can't define it on just topological spaces without extra structure
 
Gotcha. It's interesting that compactness -> completeness
 
For example, (0,1) is bounded and R is not, but they're homeomorphic
 
4:01 AM
Right
 
Zee
Compact is complete only in metric spaces
 
Well, "complete" is only defined on metric spaces
 
How can you talk about convergence outside of metric spaces?
 
and ordered spaces as well actually
Dedekind completeness, the least upper bound thing
 
4:03 AM
@orbit-stabilizer I think you can talk about convergence in any topological space, Munkres does the same.
 
@orbit-stabilizer A sequence $(x_n)$ converges to $x$ if every open set containing $x$ contains infinitely many elements of that sequence
 
Okay that works. But you can't have cauchy sequences though, right
 
"Completeness" isn't a topological property (take (0,1) and R again)
 
Or @Akiva, every open set of $x$ contains an element of that sequence. Same thing
 
Oh, actually I think we're both wrong
 
4:04 AM
Really?
 
It needs to contain all but finitely many elements of the sequence
 
Yeah, I think mine covers that
 
This way, (0,1,0,1,0,1,…) doesn't converge to anything, like it's supposed to
 
Oh
I see
You're right.
 
The other definitions would have it converge to both 0 and 1
 
4:05 AM
Man, the subtleties
 
But yeah, as long as you give the compact set a metric, it will be complete and totally bounded
 
Why we can't define cauchy sequence in arbitrary topological space @AkivaWeinberger
?
 
Zee
Look into Nets and filters
 
In R, every Cauchy sequence converges. In (0,1), not every Cauchy sequence converges. But they're homeomorphic
In addition, "uniform continuity" doesn't make sense on topological spaces either
 
That's not really getting at the heart of the question though
 
4:09 AM
1,1/2,1/4,1/8,1/16,1/32,... converge to 0, but 0 does not exist in (0,1)
 
Yes, I think we understand that
It's more: let's try to define Cauchy sequences without a topology, what doesn't work
without a metric**
 
Without a metric, how can we have a notion of consecutive difference, which is a key property of Cauchy sequences?
 
Same thing that goes wrong when you try to define uniform continuity, another thing that can't be defined on topological spaces
Essentially, the problem is the "uniformly" bit
 
We can still have a sequence of points, can we not? @Secret
 
but a Cauchy sequence is one where the difference between terms will vanish as the term number gets large?
So it is stronger than the notion of a converging sequence?
 
4:13 AM
Yeah
No
Every convergent sequence is cauchy, but the converse is not always true
 
@Secret In (0,1), the sequence you gave (2^(-n)) is Cauchy even though it doesn't converge to anything in the set
 
I think. I could ble wrong
 
how do you solve for the second value of t for 0 = t(25-4.9t) . i know t = 0 but the other value am unsure
 
To take the nonstandard analysis perspective: we can ask whether something in $X^*$ is infinitely close to something in $X$, but we can't always ask whether something on $X^*$ is infinitely close to $X^*$
 
if $a\cdot b = 0$, then $a=0$ or $b=0$. This is not the exclusive or. @WDUK
 
4:15 AM
The topology has no idea that $N\not\approx N+1$ but that $N\approx N+\frac1N$
 
so -5.1 i get roughly
 
@WDUK, seems right
 
since 25 = -4.9t
 
See:
In the mathematical field of topology, a uniform space is a set with a uniform structure. Uniform spaces are topological spaces with additional structure that is used to define uniform properties such as completeness, uniform continuity and uniform convergence. The conceptual difference between uniform and topological structures is that, in a uniform space, one can formalize certain notions of relative closeness and closeness of points. In other words, ideas like "x is closer to a than y is to b" make sense in uniform spaces. By comparison, in a general topological space, given sets A,B it ...
I once looked at these but then forgot most of it
 
4:17 AM
@AkivaWeinberger, hmmmmm I see what you're saying. Here, X* and X are disjoint?
 
but I think the idea is that these have an intermediate amount of structure between topological spaces and metric spaces
and they let you define uniform continuity and completeness
@orbit-stabilizer No, they're nonstandard analysis thingies, it would take me too long to explain them
I guess I shouldn't have mentioned them
 
@orbit-stabilizer thing is in my book regarding a falling object it seems to not calculate it that way... see image i.imgur.com/rdQX3Cq.png why do they switch it to a positive surely since its -5.1 its not a real solution and should be disregarded?
 
Essentially, $X^*$ is $X$ "with infinitesimals added"
 
Haha yeah, I don't know non-standard analysis. Uniform spaces look interesting though.
 
and $N\in\Bbb R^*$ was taken to be an infinitely large element of $\Bbb R^*$
 
4:19 AM
@WDUK, my bad. You calculated it wrong.
 
(Nonstandard analysis is unrelated to the uniform spaces thing)
 
I found this article relevant to our discussion.
 
oh ?
 
$25-4.9t = 0 \iff 25 = 4.9t \iff t = 25/4.9$ which implies that $t > 0$.
 
Is a "linear space" an old-timey name for a vector space? @Silent
 
4:20 AM
ah okay
i see it now
thank you
 
@AkivaWeinberger Yes
 
@AkivaWeinberger, I think so. I saw it in my calculus of variations book as well
 
It's a better name than vector space
 
It does make sense, doesn't it
 
it's a Bourbaki word, hence a good word
 
4:21 AM
Hah
 
"A vector is an element of a linear space".

Doesn't have that same ring to it though.
 
I don't know what makes something a "space", though. What class could have both vector spaces and topological spaces as subclasses?
 
There doesn't need to be a parent class
 
I guess they're both generalizations of $\Bbb R^n$
 
@AkivaWeinberger the category $\mathsf{Space}$
 
4:24 AM
Thanks.
 
You are welcome
 
I don't think there are any conditions on something being a space. I think it's pretty arbitrary.
The dual space to the dual space is called the duel space where all the yugioh battles occur.
 
Not to be confused with Atari's Space Duel
 
that's the double dual
 
@TedShifrin I was thinking if the straw has width, i.e. its walls enclose an interior.
 
4:27 AM
(dual space)' = space duel
 
Double Duel is not a bad name for a video game
Thing is I think you'd need four players
 
it's like a double date but for losers
 
There's something called the dual numbers, called that 'cause… I dunno
It's the ring of things of the form $\{a+b\epsilon:a,b\in\Bbb R\}$, where $\epsilon^2=0$
The weird thing is that, if $f$ is a polynomial, then $f(a+\epsilon)=f(a)+f'(a)\epsilon$
Same with rational functions actually, so long as they're defined at the input.
 
$\epsilon \not \in \mathbb{R}$?
 
Yeah it's just an extra thing
 
4:31 AM
Is this from nonstandard?
 
Nah
It's just a symbol, where we're giving it this rule of squaring to zero
 
How can i find a linear fractional transformation which maps 0,i,-i to 1,-1,0
 
If you know about quotienting rings, this is $\Bbb R[x]/\langle x^2\rangle$
or $\Bbb R[\epsilon]/\langle\epsilon^2\rangle$ I guess
 
I know about quotienting groups :P
 
Affine space is very different from vector space as far I recall
 
4:34 AM
yeah
 
It's essentially a hyperplane in a vector space of one dimension higher, that doesn't go through the origin
 
I know that for a linear transformation (Tz1,Tz2,Tz3,Tz4) = (z1,z2,z3,z4) where the R.H.S is the cross ratio.
But i cant find such T
 
Yeah, lacking the zero vector actually change some of its properties
 
@Secret, hows your phd going
 
@Shobhit Honestly I would do this by trial and error.
$iz$ maps 0,i,-i into 0,-1,1
 
4:36 AM
I am currently analysing my second batch of data, there is a trend slightly different but nothing interesting yet. I am trying to get to the 3rd set of data which I knew something interesting happened there besides a trend
 
What sort of data is it?
 
$2iz+1$ maps them into $1,-1,3$
$-2iz+1$ maps them into $-1,1,-3$
$\frac1{-2iz+1}$ maps them into $-1,1,-\frac13$
 
One thing that is clear, though, one of the ligands seemed to control the whole set of trend while the other two just modulates it. The data are molecule geometries and energy. Since my molecule has 3 ligands to vary, and there are 3 important geometry patameters each, it is at least 6 dimensional data
I am planning to get the trends analyse it before finding some hints on the explanation of the trend by analysing their electron density
 
Interesting. How are you going about analyzing it? If you're talking about a trend, is there a time component?
 
If $a:=-2iz+1$ and $b:=\frac1{-2iz+1}$ then $a-9b$ maps them into
$8,-8,0$
 
4:40 AM
Currently I make scatter plots with colour gradient which give me how two of the parameters are correlated (because the literature said they are correlated). This allow me to extract the ordering of those effect
 
So then an eighth of that
 
Two different plots with the colour gradient in two directions will allow me to visualise 4 dimensions at a time
 
Probably not the best method, I'll admit
 
and no they are all ground states molecules, thus no time component stuff yet
 
It's interesting how covariances are everywhere. Surprising how many things are linearly correlated. Or can be modeled as such.
What area of chemistry is this?
 
4:42 AM
Oh, wait, are Möbius transformations closed under addition?
Dammit, I think they aren't
 
Organometallic and computational chemistry
 
I do like what you did, unfortunately i wont be able to do it, nor might the professor will award me marks for that. Any algebraic method? @AkivaWeinberger
 
Catalysis sector
 
wow, sounds cool
I cannot do chemistry for the life of me. I hate it with a passion.
 
@Shobhit I think the $a-9b$ step doesn't actually leave you with a linear fractional transformation anyway, sorry
 
4:43 AM
I think I got scarred by the first year chem labs.
 
what is the notation for when $\mathcal A$ is a subcategory of $\mathcal B$?
 
Oh ok
 
@APerson I think just $\subset$?
 
im so tired of school
can't wait for summer
 
@AkivaWeinberger hm, ty - I think many people just write it out too, but I thought there must be some alternative
 
4:46 AM
@orbit-stabilizer I hate the summer
I LOVE the snow
 
$\mathcal A <_{C} \mathcal B$
 
I'll prolly settle somewhere in Siberia if given the chance
 
driving in the snow is not fun
 
not if you get the relevant tires
 
Still. Traffic just runs slower.
 
4:48 AM
@Shobhit New try: We have 0,i,-i. We want 1,-1,0
$z+i$ gives $i,2i,0$. $\frac{z+i}i$ gives $1,2,0$.
 
also, a coreflection arrow is a homomorphism of the supercategory right?
 
$\frac i{z+i}$ gives $1,\frac12,\infty$.
$\frac{2i}{z+i}$ gives $2,1,\infty$.
$\frac{4i}{z+i}$ gives $4,2,\infty$.
 
$\frac{4i}{z+i}-3$ gives $1,-1,\infty$.
 
Yeah, I understand. I live in Canada :P
 
4:50 AM
Simplifying, $\frac{-3z+i}{z+i}$ gives $1,-1,\infty$.
Inverting, $\frac{z+i}{-3z+i}$ gives $1,-1,0$.
 
@orbit-stabilizer I see. I'd love going there some time. But I'd not want to live there if given the chance
 
How in the world did you come up with that? @AkivaWeinberger
 
Haha, why? I think it's amazing.
 
Also equals $\dfrac{1-iz}{1+3iz}$
 
it's a beautiful country no doubt, but it has something err... wrong with it. I prefer not to mention those things here at the risk of getting banned from here
 
4:53 AM
I can read what you did, but i wont be able to come up with something like that. Its very cool. But i have to do this in my exam, how can i find the transformation algebraically
@AkivaWeinberger
 
Compared to the states, I like our healthcare and education :P
 
I suppose I'm just not liberal enough for living in Canada, in a nutshell
 
I guess I was helped a lot by the fact that one of the targets was 0
 
Oh. The political correctness stuff? It gets overblown. A lot of people here aren't like that.
 
I don't know if this sort of thing would work in all cases
 
4:55 AM
@orbit-stabilizer What the general population thinks doesn't matter much, only what the law-makers do, who only listen to those whose voices are the loudest, who in turn aren't very moderate folks
 
Yes, there may be a case where such a transformation may not exist, and we can be stuck on this method for 5-10 minutes manipulating ratios XD @AkivaWeinberger
 
Strange phenomenon no 1: it seems that everytime I recently acquired a new interest on something to explore its knowledge, for the coming 2 months the frequency of seeing discussions involving that thing increases above average levels
 
I guess. I'm just speaking personally though - I haven't felt the effects of these things
@Secret there's a term for that
Baader-Meinhof phenomenon
 
aka glitch in the matrix
agent smith
 
@orbit-stabilizer You haven't experienced its effects because there's only a probability, which is certainly a lot less than 1, but you becoming the subject of the abuse of perverted laws does carry a probability much higher than what it should be
 

« first day (2777 days earlier)      last day (2261 days later) »