« first day (682 days earlier)      last day (4635 days later) » 

07:00
@Eugene That's on computing the derivative. However that is not really what my question is abou
anon is answering my question I believe
i guess i don't really understand your question. sorry.
@anon Are you applying the inverse function theorem?
As you wrote it, I believe I am.
@anon Sorry I meant the implicit function theorem
sorry
wait
do I apply the inverse or implicit function theorem to my problem??
@Eugene I believe there are two sites: computer science - cs.SE (which is in beta) and theoretical computer science - cstheory.SE. One of them is for research-level question, one is more-or-less on MSE level. I believe the lower level is at cs.SE.
07:02
@anon Now I'm confused whether to apply the inverse function theorem or implicit function theorem
@BenjaminLim Just a guess, is your problem to use the implicit function theorem to prove the inverse function theorem? I think that's what I sketched.
Well not really.
But I think I can use it to prove my problem.
Because my $u,v,w$ are actually
@MartinSleziak i've seen some of the answers on SO and cs.se. almost all of them recommend cormen. however i wish to go a bit further.
$u = x + yz$, $v= 2e^x\sin z + y^2 $ , $w = xyz + y$
@anon I am trying to see if I can solve for $x,y,z$ around $(0,0,0)$
@anon Do I use the implicit or inverse function theorem?
07:05
@BenjaminLim this looks like the example on page 227 though...
@Eugene Ok let's look at that
We know the existence of a mapping $g$ defined about $(3,2,7)$ such that $g(3,2,7) = (0,1)$ and $f(g(y),y) = 0$
You want to invert a function around (0,0,0) (i the codomain), so that's inverse function theorem.
@Eugene What I am curious to know is what is this $g$
@Eugene I just wanted to make sure that you know that there are two sites, because of your comment here. I am not saying that some of those sites would be better than MSE.
@anon ah ok!! No wonder I could not apply the implicit function theorem, in there the function must have a different domain from codomain
07:08
@BenjaminLim also this is similar to question 19 on page 241
Note that the inverse function theorem, as the sketch I gave highlights, is sort of a special case of the implicit function theorem. The implicit function theorem allows you to solve for one set of variables in terms of another set of variables, when both sets of variables are all related to each other in a complicated implicit equation.
@MartinSleziak oh i know about them both. i was responding to caozhu's remark.
@anon Yes I am beginning to see it now!!!!
yes!!!!!
@Eugene let me do that problem now
Of course (U,V,W)=(u(x,y,z),v(x,y,z),w(x,y,z)) is a rather explicit equation.
@BenjaminLim also this is the solution for 19 if you need it
07:09
@anon What would be the inverse if I wanted to compute it?
@Eugene Perhaps you should explain in your question that you want something more/something different than Cormen et al. Otherwise you probably receive it as an answer.
@Eugene No homo I love you man XOXOXOXOXOXO
Sorry for the interruption, this was the last one concerning Eugene's algorithms question.
:4983233? I'm confused again...
@MartinSleziak i thought i said i was looking for something a little more. i guess i should make it explicit. thanks for the suggestion.
"no homo" is what straight dudes say when they want to express affection to another man but feel the need to point out that no, they're not gay.
07:12
:4983236
@anon ah. it's so funny how there are so many subtleties in language used in foreign countries. even after 5 years in north america i don't get it.
@Eugene hahahahahahahhaha
Don't come to australia, they speak a lot of slang here
@Eugene Do you know what it means to get maggot?
@BenjaminLim my cousins live in melbourne. also i know a guy from canberra here in waterloo.
@BenjaminLim I have no idea how to invert that system.
@anon yeah the inverse function theorem doesn't really tell you how to compute the inverse....
07:14
I went through most of middle school not realizing my last name is a sexual innuendo. I even made an email address out of the nickname I was given, totally oblivious.
@anon what is your last name? You can delete the comment afterwards
Nah.
@anon i saw a conan interview exactly like that. the artist's name was kate micucci
her live show was called (unfortunately) playing with micucci
07:16
hah hah
@Eugene Let us call the first equation $f_1$
the second $f_2$
and the third $f_3$
@Eugene that exercise in rudin
ugh. are you going to make me DO analysis?
then form the function
because i switched to number theory to avoid this!
$f = (f_1,f_2,f_3)$ that goes from $\Bbb{R}^4$ to $\Bbb{R}^3$
07:17
actually, solving the system set equal to zero (rather than inverting it in general) sounds possibly doable
lord i hate analysis...
@Eugene here is the matrix for the derivative of $f$
ok... bring it on!
$\left[\begin{array}{cccc} 3 & 1 & -1 & 2u \\ 1 & -1& 2 & 1 \\ 2 & 2 & -3 & 2 \end{array}\right]$
07:19
If we look at the $3 \times 3$ block without the $u$
its determinant is non-zero
everywhere
yup. i trust you...
so the implicit function theorem asserts the existence of a function $g$ defined everywhere
and $f( g(u),u) = 0$
yes...
oh I messed up the question in rudin
yes.
07:21
we actually want to solve for $z$
not $u$
the determinant of the matrix you gave me is $0$
my trust in you was misplaced...
yes
Whoops sorry man
I got -3 + 7 - 4 = 0 at the end
face palm
sorry man
I misread the question
lol.
Ok
let's now look at the matrix formed using columns 1,2, and 4
yes.
07:23
the determinant is
not zero
yup
in that neighborhood
what neighbourhood
the function $f$ is zero everywhere @Eugene
@BenjaminLim right whoops.
07:25
@Eugene The inverse function theorem asserts tthe existence of a function $g$ defined everywhere such that $f(g(z),z) = 0$
yup
@Eugene How does this mean that the variables $x,y,u$ can be solved for $z$??
that's the question in rudin yes?
yes
read page 225
07:26
I am seeing the page
but
I don't get how this means from the existence of $g$
you don't see why?
@Eugene No
I don't see how the existence of $g$ such that $f(g(z),z) = 0$ means we can solve $x,y,u$ in terms of $z$
it's just due to the invertibility
ah crap how???
invertibility of what?
the matrix $A_{x,y,u}$
07:28
ok
I get that yes
the matrix is invertible I calculated the determinant was not zero
so if the matrix is invertible then we can just multiply by it's inverse
multiply what by its inverse?
@anon i might need you to translate for me...
You see I don't get this business for solving in $x,y,u$ for $z$?
@Eugene what's up?
i'm trying to explain that the invertibility of the matrix allows $x,y,u$ to be solved in terms of $z$.
read this and the message I posted before
@Eugene I sort of see how
basically
the matrix of partials wrt x,y, u?
but not really you know what I mean?
07:32
$A_{x,y,u} [x y u] (column) = (z -2z 3z)$
so since $A_{x,y,u}$ is invertible just by multiplying by the inverse we get $x,y,u$ in terms of $z$.
There's going to be a geometric explanation of why the derivative being nonsingular allows a local inverse, but I can't think of it atm.
@Eugene What is the right hand side of the linear system?
just the $z$'s
column vector
07:34
@Eugene Where did you get this?
@anon yah essentially
@anon just math?
this is question 19 in rudin
what is $(z - 2z3z)$?
he means (z-2,z,3z)
no i mean $(z, -2z, 3z)$.
@Eugene Ah crap!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
07:35
?
the matrix of partials $A_{xyu}$
i'm confused again...
Is just
yes?
the coefficients
07:36
hahahah
of $x,y,u$!!!!!!!!!
exactly!
you got it
@anon thanks. my analysis verbiage sucks.
I didn't do anything.
@Eugene But in this case it was a nice linear system
how to do it in general?
well he seems to understand it after you came into the picture.
@BenjaminLim i think it's just the same thing.
07:37
ah crap!!!!
so @Eugene if we look at page 227
@anon now i'm the confused one. i don't understand what is going on with all these ah craps.
@Eugene look at page 227
@BenjaminLim yes?
ah yes.
that's what you're looking for
I want to solve $x_1$ and $x_2$ in terms of $y_1,y_2,y_3$
which is what i was trying to tell you.
07:39
I know this can be done
near the point $(0,1,3,2,7)$
exactly!
i was trying to tell you to look at 227 for your answer...
@Eugene It was not until you wrote out the linear system Ax = b that I saw this shit
Hmm. Is $\mathrm{ord}_p(H_n)=\lfloor \log_p n\rfloor$?
@anon what is $H_n$?
@BenjaminLim ah yes. that's it.
harmonic number, 1+1/2+1/3+...+1/n.
07:41
how do I set up the linear system for the one on 227
@anon sounds right. not sure though.
@BenjaminLim the example goes through the whole thing. read it first.
no it computes the derivative of $g$
@Eugene Do I want to solve $y_1,y_2,y_3$ for $x_1,x_2$ or the other way round?
I think it is the other way round
in this case it's the other way round since we can't tell if a 2x3 matrix is invertible.
ok
So I have the system
$2e^{x_1} + x_2y_1 = 4y_2 - 3$
and $x_2\cos x_1 - 6x_1 = 2y_1 - y_3$
yes
@Eugene How to write a matrix there's like $e^{x_1}$ and $\cos x_1$ in there?
@Eugene
i'm thinking... it looks like it's a consequence of theorem 9.27. i only know existence though. i'm trying to think of how to compute the solution
07:51
exactly
that's why I said it's not always the case we have a nice shiny linear system
aren't you just trying to establish existence though in your homework?
well I am actually studying for an exam
and I just thought it would help in my understanding
I will ask a question on the main site
yah the nonlinear case seems tough to explicitly derive a solution. sorry i couldn't be more helpful.
no worries
eugene you already made me see some of the light man!!
@BenjaminLim hey this might help you
also i think it might not be possible to find the explicit solution but only establish existence.
@BenjaminLim anyway sorry i couldn't help more. good luck with your exam.
08:12
I stick to the notation $a_{m,n}$ instead of the ambigious notation $a_mn$.
a_{mn}
$a_{mn}$
wow. it's 4. goodnight all then!
Good night.
 
6 hours later…
14:42
Where's everybody?
15:32
@Clash It's Saturday morning!
Mmmmh.. can wolframAlpha do an error?
@Dylan, up for a simple eigenvalue question?
Well, give it a shot.
Thanks! Ok, so I asked it here math.stackexchange.com/questions/158774/…. a) and b) I already have answered. I believe c) and d) should be easily extracted from b), but I'm missing something. Martin is trying to help me at his comments but I still don't get it
Simple?
Hm. It's a bit long. I'll try to look later.
15:46
Sorry, I have no idea about difficulty! For me everything is hard, but I thought it was simple because Martin made it sound like so. It's just that I thought that having b) answered should make c) easy?
Maybe it's not hard — it's not that I don't recognize the notation — but I don't know anything about numerical methods.
Does anyone know how to type Matrices in $\LaTex$?
doesn't seem to work
This $$\left| {\matrix
{{c_0}} & {{c_1}t} & {{c_{ & 2}}{t^2}} \\
0 & {{c_1}} & {2{c_2}t} \\
0 & 0 & {2{c_2}} \\

\endmatrix } \right|$$
@PeterTamaroff \[ \begin{vmatrix} a & b \\ c & d \end{vmatrix} \] seems to be easiest way. Sometimes the spacing is not so good and you have to muck around with array, though.
@DylanMoreland Ok. Works fine, though! Thanks!
 
1 hour later…
17:08
This is interesting... I have been here since early this morning, but I don't see myself in the gravatar list
Ah, there I am :-) my gravatar just appeared.
Next to me!
@MatsGranvik Yes, yours floated in just before I spoke, but mine seemed to pop in.
very odd.
@DylanMoreland There are also the \pmatrix (parentheses) and \bmatrix (brackets) environments for matrices.
17:46
Heya ^^
Hi everyone. I'm going to be without internet for 10 days starting tomorrow. Just so you know.
Also without books.
The books will be replaced with ponies.
have a fun trip in the desert
So by the time you get back I'll be gone :(
17:57
Thanks : )
Hey guys =)
18:36
Hi people
@PeterTamaroff: what request by Asaf?
@PeterTamaroff: How about f(1)=1/2 and f(n)=f(n-1)^(1/n) from there? I think that is both increasing and concave, but $f(n)^n$ increases too.
18:58
@HenningMakholm transfinite induction is under ordinals apparently
@HenningMakholm I really don't care now =P but thanks anyways.
@Eugene That makes sense, I think.
so that's settled then @HenningMakholm
19:22
@HenningMakholm I guess it only holds for certain homographic functions.
I see we've had some drive-by starring.
@anon What?
Utterly random comments are strewed across the starboard.
3
Indeed.
@anon Ha, sure.
Do you find that worrying?
19:24
No. It happens on occasion, like chance weather.
@anon Better than strewing them across the port.
There, it just happened again. Who would star "Indeed"?
@robjohn That colors in posts be avoided, he has an eye condition.
@skullpatrol [whistles]
Busted :D
There is a warm wind blowing the stars around tonite...
@MadScientist: Hi, I flagged it. Is it possible to remove that nonsense?
@Gigili It's neither offensive nor spam, I declared the flag invalid
@MadScientist You say so? I'll email the team then when you're not capable of doing it.
Let other mods decide when you cannot, there's no force.
@Gigili It's a comment that is certainly critical of you and attacks your actions. It's not on the level that a spam/offensive flag would be valid. For general disagreement I'd recommend using the ignore user feature
@MadScientist How's the "Beta test" going for Biology?
@skullpatrol It's pretty quiet
19:42
@MadScientist That's not good.
It's not too quiet
Most beta sites take a while to take off
@MadScientist Could you remind me what you were doing again, it seems somewhat interesting =)
@N3buchadnezzar What do you mean exactly?
I'm usually not in this chat, unless I see some flags ;-)
Oh right, I see. There is a new biology exchange site coming up ?
It has launched for a while now
19:44
=)
[For the record I declared "Not Sure" (with the description "no strong opinion") on the flag. I'm surprised four people starred the comment given the unfair accusation against Gigili in it, and I'm not really comfortable with the comparison.]
@anon We don't want you to feel "uncomfortable" here ;-)
Is anyone here watching Hellsing OVA`s ?
I don't watch Hellsing. I am waiting for the new chapter of Liar Game to come out, on the other hand...
@anon I'm waiting for the next episode of True Blood!
BTW, what is the fuzz between Gigili and anon?
19:52
@anon That manga is genius!
@anon I truly enjoyed the musical chairs, and the first round.
@PeterTamaroff There is no fuzz, Ragib simply made a comparison (or a contrast rather) between my suspension and her incorrect help to Jordan for differential equations. Gigili and I are amiable.
@anon Oh, cool then.
@N3buchadnezzar Yes, musical chairs was fun. I did not like the very first round of the manga, it was so predictable and the characters were 0-dimensional. (Not like they've improved much, but the games got more interesting.)
@PeterTamaroff It has 5 stars on it----->>>
@skullpatrol I read it but I couldn't follow what was going on.
19:56
In all likelihood because skull just put the fifth one there to spite me a few minutes ago. :)
I think a new book just got scanlated though, according to my manga reading list I have read 137 chapters. Not sure what the latest it though
i hate people who put about 1000 tags on a single question
@Eugene Hahahaha such as?
@anon No, that was me! I do not resonate particularly well with Gigil either.
user33836 apparently
19:57
@anon Come on pal, I would never spite you. I'll remove it then :D
Ah. It just happened that the fifth star came right around the time of this comment.
does the number of stars really matter though?
@Eugene Link to the question?
Hi guys!
it's on the front page. the combinatorics one. i edited it cause it annoyed me
20:02
@Eugene It does if you are a star counter ;-)
Could someone explain me why I can't evaluate this integral:
$$\int_0^1{\frac{e^{2x}+2e^x+1}{e^{2x}+e^x+1}}\,dx$$
Doing:
$$\int{\frac{N(x)}{D(x)}}\,dx = \int{Q(x)*D(x)}\,dx + \int{R(x)}\,dx$$?
If ya'll don't like a starred post, just star a few other messages to get rid of it.
@unNaturhal explain me?
@unNaturhal who knows why anything in math works. it's all voodoo anyway.
@unNaturhal What are Q,D and R supposed to be then ?
@Eugene Arturo does.
20:05
Apparently N=QD+R, but he forgot you need to divide N by D so you should get int Q plus int R/D ...
@N3buchadnezzar N(x): Numerator, D(x): Denominator, Q(x): Quotient, R(x): Rest
@unNaturhal Yes, but you have to find those too ;)
@unNaturhal Remainder! =)
@PeterTamaroff Sorry...
20:06
Anyway, I would start by setting $u=e^x$, then I would do, $\cdots$ ,then poff! The answer.
With $u=e^x$ you can write $\displaystyle\frac{e^x}{e^{2x}+e^x+1}=\frac{u}{(u+1)^2-u}...$
Note that $$(e^x+1)^2=e^{2x}+2e^x +1$$
And that
$$(e^x+1)^2-e^x = e^{2x}+e^x+1$$
@N3buchadnezzar I got it. I made the division, and I obtained $Q(x) = 1$, $R(x) = e^x$
So that
$$\frac{(e^x+1)^2}{(e^x+1)^2-e^x}=\frac{(e^x+1)^2-e^x+e^x}{(e^x+1)^2-e^x}$$
"The result follows at once." =P
the downvoter is back!
20:08
@unNaturhal Then you should get the same as what you started with, if you put on a common divisor right? Try that! The only problem is that e^x/(D(x)) is harder to evaluate than your original integral.
leo
leo
@Eugene is it?
downvoters, those rascally fiends
yup. whoever it is has been going through my questions and answers and downvoting without explanation. about 3-4 a day so it doesn't get reversed due to serial downvoting.
@PeterTamaroff Yeah I thought that it's possibile to solve in this way, but I can't understood why the division doesn't work...
@N3buchadnezzar I'm sorry, but I haven't understood..
@unNaturhal But it does work!
20:12
You end up with $$1 + \frac{e^x}{(e^x+1)^2-e^x}$$ but that is not much easier to integrate than your original expression! (In my eyes)
leo
leo
@Eugene have you consider start a meta thread?
Doing the long division takes longer than a few simple substitutions.
there are many meta threads on serial downvoters
@leo tamaroff asked me to do that too. its a hassle though...
@N3buchadnezzar or you could just notice the numerator is one $e^x$ more than the denominator. hardly "long" division..
20:13
@PeterTamaroff @N3buchadnezzar Wait, I write the steps that I followed..
@leo There is not much a meta post can achieve, the community can't do anything about serial downvoting. This is better handled by the automatic vote fraud script, or in extreme cases by moderators or SE employees directly.
leo
leo
@MadScientist I think moderators must do something about
@Eugene don't was me =)
see you !
@leo Meta posts are in part about involving mods, but they are mostly about involving the community.
@Eugene Eugene, doesn't $$(f \star f^{-1})(n) = I(n)$$ imply $$\sum_{d \mid n} f (n/d) f^{-1}(d)=1$$?
And moderators are very limited in what they can do about serial downvoting
20:16
what the crap is star?
@Eugene Dirichlet multiplication.
$$Q(x) = (e^{2x}+2e^x+1) \div (e^{2x}+e^x+1) = 1$$
$$R(x) = e^x$$
$$\int{\frac{e^{2x}+2e^x+1}{e^{2x}+e^x+1}}\,dx = \int{e^{2x}+e^x+1}\,dx + \int{e^x}\,dx$$
i have no idea.
@Eugene Dirichlet convolution? Product of arithmetical functions.
ask anon
20:17
Oh, wait.
I think I got it
The star is convolution of two functions.
(asterick)
@anon I got it. Nevermind. But I'll ask about other stuff, maybe later!
I dig a pygmy, by Charles Hawtrey and the Deaf Aides. Phase one, in which Doris gets her oats.
@unNaturhal Do you realize you've just written $$\int \frac{N}{D}=\int\frac{QD+R}{D}=\int Q+\int R~?$$ In other words, you have $\frac{QD+R}{D} = Q+R$. You sure you aren't forgetting a little something something?
@PeterTamaroff No, the Dirichlet inverse of a function means that $f*f^{-1}=\mathrm{Id}$, meaning that $\sum_{d|n}f(n/d)f^{-1}(d)=\color{Red}\delta_n$. (This is not the compositional functional inverse.)
@anon But $\operatorname{Id}$ is $[1/n]$
@anon Oh, now we're good. =(|)
20:23
Sorry, I'm using the wrong sense of Id. What I mean is the identity under convolution, which is the kronecker function (1 if n=1, 0 otherwise)
@anon Yes, yes. =)
Also what does [1/n] mean?
Perhaps the Iverson bracket [n=1] for the delta?
@anon Damn!!! You're right!! Thanks, I try to correct :/
@anon The nearest integer. It is $1$ for $n=1$, then $0$ for $1/n$ is fractional $<1$.
ie the floor function
20:28
oh. that's a complicated way to write it.
@anon I do prefer Iverson brackets.
Anyone want to see this?
@skullpatrol Who did that? Is it a real graph?
ok i'm going to do some real life work now like asaf. bye all!
@PeterTamaroff It represents the connections between "friends" on Facebook.
20:31
I would have never guessed.
@skullpatrol World wide? Based on real data?
@PeterTamaroff according to Facebook...
Notice how few people use FB in China.
20:46
italy is so much brighter than spain. and indonesia is just as saturated as the eastern us and much of europe.
I can't find the "friend capital" of the world?

« first day (682 days earlier)      last day (4635 days later) »