« first day (1790 days earlier)      last day (3235 days later) » 

12:11 AM
@TedShifrin your lecture on the triangle in equality is fantastic. I have never understood it that well.
 
@r9m what was the first one?
 
r9m
@robjohn the one she mentioned above :) see here
I was gonna put those on my blog .. but oh well .. someone will now answer that and hog a lot of reputation :P
 
@TedShifrin I knew these were important, but we brushed over it too quickly. I didn't have time to soak it in. Now I see how one establishes this stuff for the non 2D case. Wow, and it relates to calculus!
 
r9m
@Chris'ssistheartist seems the version with $H_n$ comes with $\psi^{(1)}\left(\frac{1}{3}\right),\psi^{(1)}\left(\frac{2}{3}\right)$ and $H_n^{(2)}$ comes with $\psi^{(2)}\left(\frac{1}{3}\right),\psi^{(2)}\left(\frac{1}{3}\right)$ .. Interesting indeed :D
 
I wish I found it that easy too
 
r9m
12:21 AM
@robjohn the original context was this :-) .. I asked if we could determine the closed form of the version with $H_n^{(2)}$ as well :)
 
@robjohn
 
@PedroTamaroff yes?
@r9m I'll have to take a look at that... interesting
 
Is it true that for two continuous random variables, $$P(Y\leqslant t)=\int_{\Bbb R}P(Y\leqslant t\mid X=s)f_X(s)ds$$ where $f_X$ is the density function of $X$?
 
@PedroTamaroff I think you might need independence of $Y$ and $X$
 
Looks shady, since $P(X=s)=0$ for every $s$.
 
12:25 AM
@PedroTamaroff $P$ would have to be a density...
 
My problem is the following: $X$ is uniformly distributed over $[0,20]$, and $Y$ is uniformly distributed over $[X,60]$.
I want to calculate ${\Bbb E}Y$.
I was trying to do it by calculating the variable ${\Bbb E}(Y\mid X)$.
 
$$\frac1{20}\int_0^{20}\frac{x+60}{2}\,\mathrm{d}x$$
 
@PedroTamaroff $E(Y)=35$ is what I think
 
OK, but how are you getting it?
 
12:30 AM
@PedroTamaroff For each $x$, $E(Y|x)=\frac{x+60}{2}$
 
What does $E(Y|x)$ mean? Does it mean $E(Y|X=x)$?
I was trying to do something of the sort.
Trying to claim that $E(Y)=\int_\Bbb R E(Y|X=x)f_X(x)dx$,
Seems like that's what you're doing.
 
@PedroTamaroff it is the density... the expected value of $Y$ given the value of $x$
@PedroTamaroff It follows from the linearity of expectation
 
@robjohn The density of what? Of $Y\mid X=x$?
Is my last integral formula true?
 
@PedroTamaroff $P(Y|X=x)=\frac{P(Y\land x\le X\le x+\Delta x)}{P(x\le X\le x+\Delta x)}$ density of $Y$
 
@robjohn OK, my formula seems to have worked.
How would you obtain ${\rm cov}(X,Y)$?
 
12:37 AM
@PedroTamaroff yes, it is pretty much what I had used.
The expected value can also be computed as $$\frac1{20}\int_0^{20}\frac1{60-x}\int_x^{60}y\,\mathrm{d}y\,\mathrm{d}x$$
 
Right.
@robjohn I am not sure how to obtain ${\rm cov}(X,Y)$.
I will try to find the distribution of $Y$ by myself, though.
 
@PedroTamaroff I have to go out for a while, then eat dinner. I will look into it when I get back
 
Hi @TedShifrin.
 
Heya mr @Pedro
 
Are you still in the mood to teach people some probability theory?
 
12:45 AM
Hmm ... Dunno. Depends if I know what you're talking about ....
 
@TedShifrin I don't know how I didn't see them before, but I'm checking out your lectures on YouTube right now. They're fantastic!
 
I want to find the distribution of $Y$, which is distributed uniformly over $[X,60]$, where $X$ is distributed uniformly over $[0,20]$.
 
Thanks, @Fargle.
 
@Fargle are you starting with 3500 or 3510?
 
3500 for now, @Cbjork, just to fill any and all gaps. I've noticed some gaping holes in my knowledge, or at least my base of intuition, by working through differential geometry.
 
12:47 AM
You should be doing a double integral, then, Pedro?
 
Should I?
=)
 
You want $P(Y\le t)$?
 
@Ted, do you have the Princeton Companion?
 
Nope ...
 
@TedShifrin Yes.
 
12:50 AM
@TedShifrin Ah, it's pretty cool, but quite expensive (and heavy), I guess.
I'm off to school, bye :)
 
Bye ...
So don't we have a uniform distribution on the obvious triangle, @Pedro?
 
@TedShifrin I don't know what the obvious triangle is.
 
@Fargle it's a fun class. My favorite problem from that class is prove or give a counterexample: Let $f: \mathbb{R}^2 \rightarrow \mathbb{R} $ a smooth function have one critical point $\vec{c}$ where $f(\vec{c})$ is a local min. $f(\vec{c})$ is a global min.
 
Si it should be a ratio of triangle areas.
 
@Cbjork That is quite a problem. I'll have to think on it for a bit.
 
12:55 AM
$\{(x,y): 0\le x\le 20, x\le y\le 60\}$?
 
@TedShifrin I'm trying to understand the setting.
$Y$, as a random variable, is a function of $X$.
 
Hmm, maybe I'm wrong. My brain isn't in this mode. Go on ...
 
Well, I am not sure what to do.
 
Do you have a textbook? The setting should be described there.
 
I still think I'm right.
You can do the conditional density function. It should come out the same.
 
12:59 AM
I think you're thinking about the joint density function.
 
I didn't emphasize this when I taught it ... Admittedly.
 
Should I try to obtain the joint density and then marginalize?
 
sorry, it's a trapezoid, not a triangle.
 
I have no idea what you're doing Ted!
 
Finding the area.
 
1:02 AM
I'm trying to use basic geometry, nothing fancy.
 
I am trying to find the density of $Y$.
 
I'll think a bit later ...
 
Morning.
 
Evening
 
@Huy I need an iPhone to compute 20% of a tip, or at least a colleague.
 
1:09 AM
Dividing by 5 is the same as doubling.
 
@skillpatrol You mean .5
 
20% = 1/5 = 0.2
 
Oh, you were talking to Mike.
Still, what you said is not what you meant. =P
 
Double and move the decimal.
To compute a 20% tip :-)
@PedroTamaroff Not too sure what 20% "of" a tip means =P
Just joking guys ;D
 
1:41 AM
@PedroTamaroff $$\frac1{20}\int_0^{20}\frac1{60-x}\int_x^{60}(x-10)(y-35)\,\mathrm{d}y\,\mathrm‌​{d}x =\frac{50}3$$
 
@robjohn What is that?
 
That should be the covariance
 
The covariance?
What's the reasoning?
 
@PedroTamaroff Just integrating $(x-\bar{x})(y-\bar{y})$ over the joint density
 
What is $\bar x$?
 
1:43 AM
@PedroTamaroff the mean of $x$
 
2:02 AM
@robjohn How did you obtain the joint density?
 
2:36 AM
Is the space of all linear operators a function space?
 
@PedroTamaroff Wouldn't it just be the conditional probability density function for $Y$ given $X=x$ times the probability density function for $X$? That would be $\frac{1}{60-x} \cdot \frac{1}{20}$.
 
2:49 AM
@PedroTamaroff $$\frac1{20}\int_0^{20}\frac1{60-x}\int_x^{60}f(x,y)\,\mathrm{d}y\,\mathrm‌​{d}x‌​$$ is the mean of $f(x,y)$ given the $X$ and $Y$ you described: $X$ uniform on $[0,20]$ and $Y$ uniform on $[x,60]$
 
3:00 AM
@PedroTamaroff $$P(a\le Y\le b|X=x)=\begin{cases}0 & a\le b\le x \\ \frac{b-x}{60-x} & a\le x\le b \\ \frac{b-a}{60-x} & x\le a\le b \end{cases}$$ Divide by $b-a$ and then let $b\to a^+$. If $a< x$ this will be $0$, and if $x\le a$ it will be $\frac{1}{60-x}$. Now multiply by the probability density associated with $X=x$ (namely $\frac{1}{20}$).
 
@anon Did I tell you I started a blog? It might contain entries that interest you.
Last one is about rep. thry.
By the way, I did find $P(Y\leqslant t\mid X=s)$.
I was silly not to differentiate.
 
hmm, your blog seems to about modules
 
Hehehe, in part.
You don't like modules?
 
well, I don't usually think in that generality
except for basic stuff
good job on the blog though
 
Thanks.
Hm, the first posts are not about module theory.
I am thinking about writing about complex analysis or group theory in what will follow.
 
3:14 AM
morning
 
Hello Mike.
I posted the Brauer Thrall thing I told you about.
 
I see
 
You can see the correct statement now.
@RandomVariable Yes, I got that.
Plus some indicator functions.
@RandomVariable I got that $f_{X,Y}(s,t)=\frac{1}{20}\frac{1}{60-s}{\bf 1}_{[s,60]}(t){\bf 1}_{[0,20]}(s)$.
 
3:30 AM
Can anyone help with basic homological algebra?: math.stackexchange.com/questions/1342870/…\
so the book is is just full of errors then?
 
@EnjoysMath The splitting maps don't go that way.
They are degree $1$ maps.
They go in the opposite direction to the differentials.
So $s_n:C_n\to C_{n+1}$.
 
So there are errors all over the book
 
Yes, probably. Google "Weibel errata."
You want that $d_ns_{n-1}d_n= d_n$.
 
I did
 
There's a big list.
You should find it.
 
3:32 AM
I checked. Or the kindle pages are way off
Thx @PedroTamaroff
 
hello all ^_^
 
Sawadee
 
@RandomVariable What limits did you get? You need that to integrate to $1$.
 
3:45 AM
@TedShifrin did you head about the new Californian trend of planting cacti?
 
@PedroTamaroff For $y$ between $x$ and $60$ and $x$ between $0$ and $20$ it does integrate to $1$, doesn't it?
 
@RandomVariable Yes, I got that.
Let me check if it does.
 
 
4 hours later…
7:57 AM
@r9m The problem reveals at least an amazing thing I wasn't aware about.
 
8:40 AM
@Alyosha No, that's the whole fun of it. $\mathcal{O}_R$ is not a PID when $R = \Bbb Q(\sqrt{-23})$
I think.
@SohamChowdhury I just gave a geometric/visual example of non-Hausdorff topology to @Remember. Look it up in the chat.
The general idea is to bring two spheres/anything infinitely closer, but not touching each other still.
Also, note that cofinite can be replaced with cocountable.
@TedShifrin The line with two origin is an ambiguous name : there's the $\Bbb R \sqcup \Bbb R$ with all the points except the origins identified.
That's non-Hausdorff too, though, but not homeomorphic to the thing you wrote down.
 
@BalarkaSen an odd map $f:S^n\to S^n$ has odd degree
have you ever proved that?
 
8:59 AM
@r9m it's great if you have a proof on a bill. I don't have yet any such a proof.
 
@iwriteonbananas it's in Hatcher.
that's actually a lemma you need for proving Borsuk-Ulam theorem
 
I was about to read the proof, when I suddenly became ill. So no, haven't read it yet.
 
@r9m This is related to the Knuth problem!!!
 
stop getting ill. :/
i think there is an easier proof
 
9:05 AM
hmm.
i think you're right
push that wretched map to RP^n-level
 
Mew
hello
 
im saying that because my prof suggested an exercise
 
I think you can do this without using homology at all.
Let me ponder.
 
Mew
Could a mathematician help with this:
0
Q: Could the solution to classical gravitational 3-body problem be periodic?

MewI understand that some solutions to the classical gravitational 3-body problem have been found to be periodic, such as the choreograph below: Could it be possible that all solutions to the 3-body problem turn out to be periodic given enough time, or has it been shown that for some solutions, t...

 
yeah agreed
consider the action of $\Bbb{Z}_2$ on $S^n$
that maps the non-trivial element to the antipodal map
my prof said to use the fact that there are no equivariant maps $S^n\to S^m$ if $n>m$ and conclude that an odd map has odd degree
but i forgot what an equivariant map was, so im looking that up right now
 
9:09 AM
i dunno what that is either
anyway :
$f : S^n \to S^n$ be an odd map.
push this to $\tilde{f} : \Bbb RP^n \to \Bbb RP^n$
 
consider $\tilde{f} \circ p : S^n \to \Bbb RP^n \to \Bbb RP^n$
 
this is trivial at the level of fundamental groups. agreed?
 
9:11 AM
so we can lift this guy to $\widetilde{\tilde{f} \circ p} : S^n \to S^n$ again.
 
as lifts are unique, this is just $p^{-1} \circ \tilde{f} \circ p$,
 
what's $g$?
ok
 
I meant $p$, the covering map $S^n \to \Bbb RP^n$
 
9:16 AM
well, that seems to be a nice dead end i've put myself in.
 
haha :D
well, to say that $f:S^n\to S^m$ is not equivariant means that $f(-x) \neq -f(x)$ i think
i got a shitty lecture now, i'll think about it then. gotta go offline, c ya
 
bye, while i rack my brains on this.
 
9:46 AM
@r9m I know how to do it now, referring to your series!!! :-) Well, I only need to write things down. And there is one more thing: it seems one need to also know to calculate
the variant with $H_n$.
 
9:59 AM
What is the Putnam? I see it often in regards with math, google.ee/search?q=Putnam+discrete+mathematics
 
10:21 AM
@RecognizeEvilasWaste Apart from being the name of a mathematician, it is the name of an American math competition
 
10:33 AM
hey @Soham. went to school in this much rain?
 
huh?
there was no rain today, I guess.
 
it was raining like mad outside.
 
weird. I saw nothing.
 
must've been a local rain.
 
hmm.
i hate it when it's overcast. I feel depressed. (yes, I am using that word correctly)
 
10:36 AM
just the opposite in here.
so, what kind of math have you been doing?
(and by math i mean real math, not story-book-math and fairy-tale-math :P)
 
lolic-maths
 
@BalarkaSen oh, point-set. yesterday I learned about Heine-Borel.
 
cool theorem.
 
when it rains and then stops raining, I sort of go into existential-crisis mode.
"what am I doing with my life?" etc
@BalarkaSen what proof do you know?
 
@SohamChowdhury i've never been through those.
 
10:40 AM
ah. lucky you.
 
@SohamChowdhury proof of Heine-Borel?
 
i'd forgotten the easier one i used to know. the only one i know right now uses Tychonoff.
 
oh.
what was the idea of the "easier" one?
 
Bolanzo-Weierstrass and a lot of analysis.
assuming Tychonoff, it's easy : just classify closed sets in R and extend this to R \times R \times .. \times R
 
10:44 AM
btw, I was pretty astounded by the def of "compact". It captures the intuitive idea in an interesting way.
 
indeed.
you're talking about the open cover definition right?
 
Bolzano* :)
@BalarkaSen there are others?
 
in R^n, sure.
sequential compactness, etc.
 
@BalarkaSen no, in general.
$\sf Top$.
 
ok, then the open cover def is the only one i know.
 
10:46 AM
ah.
 
note that sequential compactness is also a valid definition of compactness in metric spaces, btw
 
@BalarkaSen As far as I recall, replacing sequences by nets makes the sequential definition work in general
 
ok, i didn't know that.
i never cared about nets.
 
achha. take a torus and draw a circle on it, along the shorter circumference. (i.e. so that if you cut along it, you get a cylinder)
v = 1, e = 1, f = 1. wtf?
 
what's your triangulation?
what you did wasn't a valid triangulation.
 
10:48 AM
oh?
 
you need triangulations to get euler's formula work.
 
doing the same thing on a sphere works. (i.e. drawing an equator)
 
triangulations are subdivisions of spaces into small triangles.
 
i'll look valid triangulations up.
 
@Soham that's because the upper and lower halves are homeomorphic to triangles.
 
10:50 AM
oooh.
 
nods.
 
@BalarkaSen The nice thing about nets is that most statements in metric spaces that can be formulated in terms of sequences become "correct" in general topological spaces when formulated with nets
such as Hausdorff iff nets converge to at most one point
 
aren't nets some sorta set theory thing?
 
ehh. i'm not fond of sequential aspects of viewing everything, although i know that's a good way to see some things.
 
@SohamChowdhury No, they are just a generalization of sequences
but they are often mentioned in connection with filters, which have a set-theoretic aspect due to the fact they there might not be any non-principal ultra-filters unless one assumes AC
 
10:53 AM
@Balarka, what are you doing?
 
recovering from illness, mostly.
 
what this time?
 
inner ear infection (due to cold)
 
pats Balarka on the back
 
it's not pleasant to watch everything spin around you.
 
10:54 AM
:(
 
doesn't that hurt like hell ?
 
nah, it just gives you a positional vertigo
 
i had a terrible asthma attack two days back.
ah. anyway.
 
@SohamChowdhury yikes
 
I find it weird that Lee advertises his book as a grad thing, but Armstrong says his is an intro book.
 
10:55 AM
make sure you take inhalers timely.
 
@BalarkaSen monthly surprise for me.
@BalarkaSen inhalers don't work anymore. I take some sort of inhaled powder capsules.
I used to take adult doses of Seroflo at 5 or so.
achievement unlocked
 
hell.
 
yeah.
anyhoo.
 
@Hippalectryon ach. true.
I probably have a non-blocked nose for like a week every year.
 
10:58 AM
Same here
 
when you've grasped triangulations, @Soham, draw one on the torus and compute it's euler char.
 
yeah.
v = 1, e = 1, f = 2.
if you draw a circle.
 
no.
not a triangulation still
 
okay, okay, I got one.
 
i am betting you didn't.
 
11:04 AM
Tip of the day: If you want to get upvotes keep your answers short.
 
the minimal number of triangles required to triangulate the torus is 14, iirc.
 
Hatcher preface?
 
and your questions long.
 
@Soham hmm?
 
Hatcher says that about the torus in his preface.
 
11:05 AM
really?
 
@r9m I wouldn't say that you show too much mercy with the questions you ask in chat, but this time you have come with the devil unleashed. :-)))
 
I thought he said that in Ch. 2.
 
yes. he says that's why he uses "$\Delta$-complexes", whatever those are.
 
yep
much more easier.
now learn the definition of triangulation. you're mixing up triangulations are \Delta-complex structures.
 
I don't even know what those are.
@BalarkaSen okay.
 
11:07 AM
the former requires that each triangle is uniquely determined by it's edges.
while the latter doesn't.
 
in a sense, (singular) homology is about computing the number of vertices, edges, faces (and higher things) in a triangulation-independent way.
although i might be oversimplifying things.
 
this is what I had in mind: essentially, you can make a torus out of four C-shapes.
right?
 
"C-shapes" = "cylinders"?
 
11:15 AM
i don't know what you mean, then.
 
$\huge{\mathbf{C}}$
join two of these to get a punctured disk.
 
oh, i see. sure.
 
join two punctured disks and id along boundaries.
 
glue pairs by edges of that thickened C to get cylinders
then glue the cylinders through the boundaries
i get you.
 
pairs?
oh, yes.
you understand.
 
11:17 AM
right.
 
not a simplicial complex?
 
nope.
 
I think khata baar korte hobe.
 
I encourage that.
 
Anyway, there's a person coming over to fix the computer, so talk later. :)
 
11:18 AM
ok, bubye.
 
I'll work on this.
 
I have to get some rest.
 

« first day (1790 days earlier)      last day (3235 days later) »