« first day (1963 days earlier)      last day (3353 days later) » 

19:00
@Balarka: You should follow Mike's lead and be embarrassed about your hat compulsion and get back to learning mathematics.
@BalarkaSen Only questions related to the work of scientists/mathematicians are allowed.
I have stopped hat-hunting.
Why?
Too much competition??
@TedShifrin This year's hats are way too hard. But I cooked up this question, and wanted to post it, so I did. Plus, I liked a question on anon, so I'll post an answer anyway.
@ted: I haven't really bothered trying to prove the result, though, if I"m honest. (mostly because there are better things for me to work on---not that i am, but the point stands :P)
19:01
I targeted it for getting a secret hat, but now I don't care anymore. I'll just post an answer because I like it.
:) @Semiclassic
Mike has lots of competition this year.
Mike has work too. I encourage his efforts on trying to stay off SE.
though i did figure out the minimal example which I think conveys the idea
Last year was too easy.
Nobody really cared.
19:03
namely, take $F=\mathbf{x}^2$ and $G=\hat{n}\cdot \mathbf{x}$
Why is one x bold and the other blackboard bold?
@TedShifrin Btw, the question was who discovered the singular version of cup product (i.e., the usual one - multiply cochains). I know that the homological version was by Eilenberg&Zilber, the intersection was the original Poincare one. But who discovered this?
Do you know?
because i'm being weird :P
Huy
Huy
I thought you were using two different x
If so, post an answer!
Huy
Huy
19:04
that would have been great notion
i did, but didn't mean to
No, I don't know definitively.
I found something about Lefschetz, but they didn't give any reference.
Huy
Huy
references
Lefschetz was totally geometric, preceding the algebraic formalism, I believe.
Huy
Huy
19:05
I wish there were references for the top answer of why Feb consists of 28 days
cuz the answer doesn't make any sense at all
You can leave a comment.
Huy
Huy
there already is such a comment there, @Danu
You can always leave another or, if you don't think this should be an answer, flag the post.
Huy
Huy
too lazy
@Huy Where's that question?
19:08
@TedShifrin The other question on HSM I posted was for getting the explorer hat. It was about who discovered the topological proof of Nielsen-Schreier, but it turned out the answer was already in the wikipedia article on Nielsen-Schreier. What's worse that I linked it in my question :P
Thanks to Keith Conrad, he posted an answer: "look in the bottom of the link in your question"
@BalarkaSen Yeah, that was kind of ridiculous.
For that you should be banned, @Balarka :)
Why is it ban-worthy? I made a total fool of myself, sure, but not sure about ban :P
19:10
Get back to work, @Balarka.
Ok, I really should.
Winterbash's over for me.
Funny that when someone posts a question that's wrong, people will post wrong solutions :P
@Ted cis.upenn.edu/~jean/kruskal.pdf Here is something in Higman's Lemma that I really liked.
@TedShifrin People also do that when the question is correct sometimes. Link?
@DanielF: Here you go.
@Julian: I know none of this set theory :(
19:15
It isnt completely set theory.
It has some combinatorics as well
Kruskal's Theorem I recognize from graph theory.
(which I also don't know) :)
You sound as un-nice as prof here. :P
That is something I am going to re-prove after Higmans
@TedShifrin You have to admit, remembering all the $(-1)^k$ is hard.
19:17
I do?
But when I question him, he should be a bit more self-critical, especially when I've posted a counterexample :P
Huy
Huy
no, Balarka is just very sensitive during the holidays
@TedShifrin And, being German, he ought to know how to spell Leibniz.
@Balarka: I'm ordinarily very kind, but this guy is not stopping to think about whether he knows what he's doing.
@DanielF: Is he German or French?
Huy
Huy
probably German if he writes tz
I know you're kind and patient. A good deal more so than Mike. :P
19:19
less so?
I don't claim to be kind and patient ... but ...
I have said a lot of stupid things to you, and you've never reacted unkindly.
@TedShifrin Well, at least he studied in Munich, I recall his profile saying.
Huy
Huy
@DanielFischer "In den frühen Schriften anderer Autoren über ihn, aber gelegentlich auch von ihm selbst wurde sein Nachname analog zum Nachnamen seines Vaters auch „Leibnitz“ geschrieben."
19:20
I was working with a ninth grader the other day who hates everything academic (but particularly math) and only cares about football and basketball. He's not a dumb guy, either, but he just makes no effort to engage. I don't know what to do to motivate him ...
Huy
Huy
let him talk to me, I deal with those kind of guys every day
@Ted Did you know I tutor math around my area for $20?
(crazy I know)
@Huy Yes, but that's history, the accepted spelling was Leibniz long before Duden.
Huy
Huy
@DanielFischer: I don't know anything about this, I'm just an immigrant.
19:22
@TedShifrin Suggestion: Beat him up.
Can you give me advice, @Huy?
Not so crazy, Julian.
He'd pulverize me in an instant, @Balarka. Thanks for that.
But he is very friendly and likes me ... weird, I know.
Huy
Huy
@TedShifrin: not knowing anything about him: why is professionally pursuing football or basketball not an option for him? is he not good enough?
@Ted So a 15-year-old getting $20 an hour for tutoring twice a week is the norm?
Yes, @Julian.
@Huy, as a rule of thumb, one should assume the answer is no. And no way to know with someone 14-15 years old.
Huy
Huy
19:23
@Julian depends on the country, over here that'd be a bit underpaid.
No, @Julian, far from it. But you're far from the norm 15 year old.
In fact, the standard earning is 30$.
@Julian: Remember that when you tutor you should be socratic and not do the work for them. Also try to make sure to explain, rather than just give answers.
Huy
Huy
@TedShifrin: but I mean does he play in a big club or is it really just a hobby (even if a very important one)?
Wow. I searched online and I thought the student norm was $10-$15
19:24
@TedShifrin Looking at his profile, I can see with ~90% certainty that I study with him.
@Huy: He's just in high school, beginning of high school.
Oh really, @Danu? Now I feel bad for being mean ;P
Huy
Huy
@TedShifrin: in the US it's not possible to play at a club AND be at high school? sorry for the stupid questions but over here one doesn't exclude the other
@TedShifrin I don't know who it is.
But I study in Munich
@Ted I am actually a little too strict on them sometimes. But their parents are like "thats how they learn." I am like "ok."
In a degree explicitly called "theoretical and mathematical physics"
19:25
@Huy: We don't have clubs in this country.
Huy
Huy
???
you have like NBA and MLS
So he's bound to be in the same degree as me. Also, his interest in functional analysis shows that's he's probably a German student.
@Danu: He definitely should understand skew-commutativity of differential forms!!!
@TedShifrin There are many people here who care more about FA than geometry.
@Huy: You're talking about professionals, not 14-15 year olds.
19:26
In fact, I'd say >80% of students here don't care about geometry much.
Huy
Huy
major clubs usually bind very young talents already :P
The first topology class (graduate level) here has ~15 people attending.
@Danu: I'm fine with preferring functional analysis. But one should still know the basics if one answers easy question.
Huy
Huy
but I guess the answer is no then
19:26
Riemannian geometry had 6 students, 4 taking the exam.
@Huy: I have no way of knowing how good he is, but he's just a kid, and if he doesn't get his grades up, he'll get kicked off sports teams in school.
@TedShifrin Preferring FA as in not being interested in taking any geometry course.
Huy
Huy
@Danu: I'm the only one taking the exam in geometric topology over here :(
lollll
And they don't change it to an oral exam?
Huy
Huy
@Danu 99.9% of all exams after the 3rd semester are oral here
(in maths)
19:27
Ah... Okay.
Here, it's different.
@Huy Don't worry, it means everyone except you in your school are destined for hell after death.
Viewing the questions posed by Freeze_S confirms my suspicions.
I wonder who it is...
God prefers geometric topologists.
Huy
Huy
@TedShifrin: I don't know why but those kind of people seem to like me. probably because I'm still very young and they believe that I actually understand them and don't just act as if I do (what "older" teachers might do). I'd probably just talk a bit about sports with them, about the future, how grades at school matter for their future and so on, but that you probably all know too
How can someone have so many posts and so little rep? o_0
19:29
@Huy: This week was my first interaction with him. If I continue to work with him, I guess I'll have to try that, but it's not really my role in this "job."
Huy
Huy
@Danu: Personally, I much prefer oral to written exam. It's over much quicker and you get to show much more of what you know.
I only had one oral exam and it went terribly.
@Danu a) bounties, b) the questions and answers are highly idiosyncratic. He loves to use obfuscatory notation.
@Huy Only if you can do problems in your head.
Huy
Huy
@BalarkaSen: not everyone expects you to solve problems without pen and paper :P
19:31
@DanielFischer Classic German LMU students :D
@DanielF: He tells me he's going to delete his wrong answer, but now there are two downvotes.
Mathematical quantum mechanics is the One True God!
Huy
Huy
@Danu I actually believe we've had this discussion before :D
Probably.
<--- begins to fill outnumbered by physicists :P
Huy
Huy
19:32
@Danu: I guess everyone has their own preferences. :) glad that it works out for each of us since at your uni it's the other way around ?
@Huy Yeah!
@Danu Okay. I know that Bavarian is an incomprehensible mish-mash of sounds. But I thought Munich had enough immigrants to develop something resembling German?
@TedShifrin We're being driven from the PSE room ;)
Who's driving you?
@DanielFischer Hah, no.
@TedShifrin Eh, the "room culture" has just changed.
19:33
It's changed a lot here too.
It used to be less, but more "hardcore" physics focused people.
I dont' mind random discussions about nonsense.
But now the room is filled 24/7 with noise.
@TedShifrin I'm going to start on reading 6 because I don't want to be stuck at 5 anymore. Don't worry though, I'll definitely do your homework in 5. Trying to warm up.
Here it's either geometric topology or limits/series/integrals. :)
Huy
Huy
@TedShifrin: BTW, earlier this week I got a letter informing me of my first raise. ^_^
Ausgezeichnet, @Huy.
19:34
@TedShifrin Heh... Wel... I'd love to be able to contribute to the discussion with some more geometric stuff.
Well, @Danu, I'm happy to talk differentiable manifolds and geometry.
I'd really like to learn (from you, or anyone else), but it's hard to not feel like just "leeching"
It's not like I have much interesting stuff to say about things...
I've sent bananas some of my exercise sets in the past. I don't remember how far we got with that.
I sadly also don't have any time to spare on extra problems.
It's annoying how graduate school gets in the way of learning :D
19:35
I'm taking 5 courses (that's 45 EC, the recommended amount being 30), and 30 hours of class + all the homework is already too much.
Yeah, that's way too much.
Huy
Huy
that tends to be the trend though for maths and physics students :D
Ich stimme überein.
It's alright, because there are always 1-2 classes that are not that interesting.
oh, ted, a follow-up on that question I was just mumbling about. is there a version of it in the calculus-of-variations setting? i have in mind the fact that, if one wants to do Lagrangian mechanics with constraints, then if memory serves the multipliers are interpreted as the generalized force associated with that constraint
19:36
And you know... not that hard :P
I think my German is way too rusty. DanielF or one of you will correct it.
Huy
Huy
@Danu: over here I'm the only one who only does 30 :(
@Semiclassical Not quite---but close (if I recall correctly)
The generalized force definitely comes from the multiplier terms---but it's not equal to the multiplier AFAIK.
Ok, contraction mapping principle is peasy. Take contraction mapping $f: X \to X$, and define the sequence $a_n = f(a_{n-1})$ recursively, $a_0$ some pt in $X$. This is bound to converge somewhere as $f$ is contraction. That pt is the fixed pt.
19:37
i made a matlab script and it confirms my latest discovery wonder why did i not recieve any feedback or estimation (either negative or positive)
Essentially correct though.
@TedShifrin Heh, it's okay.
Works in general if $X$ is a complete metric space, I'd think.
I'm impressed by anything non-zero from an American ;D
hey @TedShifrin
@Danu: I have always taken an interest in languages.
19:38
:D
Yes, @Balarka, of course.
You ought to be European ;D
The EU may have something to say about that, @Danu :)
@TedShifrin I don't think 'tis that hard for Americans :P
@Balarka: Spend more time on the exercises, as that'll teach you some analysis. Particularly 8, 10.
19:40
We write $y = A_n^k x$. Then y is of the form $(x_k, \dots, x_n,0 , \dots , 0)$. Then $A_n y= (x_{k+1}, \dots, x_n, 0, \dots, 0)$.

So is the above sufficent?
Thanks. I'll have to read up Newton's method first.
Well, it's in the section.
Yeah, noticed.
It's a classic topic from a standard high school calculus class.
Newton's method?
19:41
nod
Not really AFAIK
But it's easy enough to understand at the high school level already.
i definitely knew about newton's method by the time i left high school, but i don't recall where i encountered it
It's just an algorithm to program into one's calculator, and everyone uses graphing calculators these days in high school courses.
Huy
Huy
isn't Heron for square roots Newton applied?
I've never tried reading up Newton's method.
19:42
It's definitely in first-semester college courses, but it's basic.
No proofs, of course.
@evinda In the right context, it is.
You'll learn it there, @Balarka, or else from the video if you need to.
Ok, I will think about it again.... @DanielFischer
Thank you!!!
@Huy: I don't know Heron for square roots. You mean the standard algorithm for computing square roots by hand?
19:43
@Dave hehe
Huy
Huy
I wonder how many more times I'll have to face this picture till Christmas
@TedShifrin, yes probably
@TedShifrin Ohh. This. I didn't even know it was called Newton's method.
@TedShifrin Aka "the Babylonian method", aka "Newton-Raphson".
That's based on $(a+b)^2 = a^2+2ab+b^2$, not Newton's method. But the Newton's method approach is very close, I think.
Newton's method = pick an initial guess for where a root is, approximate the function with the tangent line at that point, then find its root and use that as the new guess.
rinse-and-repeat as needed
19:44
I have used this once upon a time while studying calc. Probably found it out while playing with stuff.
So you'll get to prove how fast it converges, @Balarka.
Huy
Huy
@TedShifrin: This one.
What's cool about those exercises is that they give you a constructive bound on where the root will be, if you have bounds on first and second derivatives.
Hmm, that I don't know. I'll ponder.
@Dave state the restrictions :P
19:45
@Huy: Oh, that's not the algorithm I learned in elementary school.
Huy
Huy
it's the one we learn (and teach)
The one I learned was based on the binomial expansion. I'm not going to try to write it out here.
@TedShifrin elementary school :P
binomial expansion?
some advanced elementary shit :P
Yeah, @Danu, I'm old. When I was in fifth grade or so we learned to do this by hand (after learning long division).
i think my favorite example of an algorithm which is simple but deep is the arithmetic-geometric mean as a way of doing elliptic integrals.
19:46
They didn't explain why it worked. I figured that out a year or two later.
I never learned long division :D
Yes, @Semiclassic, that's cute.
Huy
Huy
can you divide polynomials?
it converges so rapidly, yet i've always found it hard to carry out the derivation of it
(there's a name I should be associating with that derivation and I can't remember it.)
OK, I'm gone for now. Work hard, everyone.
19:48
later @ted
@TedShifrin I am not convinced it'd be very fast. Start with a spiral. Choose a point on the spiral the tangent to which cuts the x-axis sufficiently away from 0.
It'd be very very slow.
@Balarka: A spiral is not the graph of a function.
there are definitely examples where Newton's method is a bad idea, but in general it works darn well
I mean, piece of the spiral.
The exercise you will work will show you why you need upper bounds on $|f'|$ and lower bounds on $|f''|$.
Given those bounds, it converges geometrically quickly.
19:49
Hm. Fun.
Bubye.
Thanks!
what I'd be curious about is whether those bounds are sharp
me too
@Danu Anything new on topology?
19:50
Hmm I remember learning the convergence rate of Newton's method.
That stuff #reks everything else basically in numerical methods, IIRC.
@BalarkaSen Classes are Mon/Wed.
So you're up to speed ;)
Still waiting to continue on SVK, I guess.
The next problem set is up though.
7 questions instead of the usual 3-4.
@DanielFischer I have also an other question.
We consider the linear mapping $T: \mathbb{R}^2 \to \mathbb{R}^2$ with $T(x_1, x_2)=(ax_1+bx_2, cx_1+dx_2), (x_1, x_2) \in \mathbb{R}^2$, where $a,b,c,d$ are given real numbers such that $|a|+|b|+|c|+|d| \neq 0$.
I want to show that the set $\{ M>0: ||T(x_1, x_2)||_2 \leq M ||(x_1, x_2)||_2 \forall (x_1, x_2) \in \mathbb{R}^2 \}$ is non-empty.

The following answer is given: $||T(x_1,x_2)||_2 \overset{\text{Cauchy-Scharz}}{\leq} \sqrt{a^2+b^2+c^2+d^2} ||(x_1,x_2)||_2 \forall (x_1,x_2) \in \mathbb{R}^2$.
SVK is for spaces $X$ with open cover $\{U, V\}$ where $V$ might not be simply connected.
I suspect an example where it wouldn't work well is generated by a cubic with an inflection point in the middle, since then the lower bound on $|f''|$ is only the trivial one
19:52
So a generalization of the stuff you did.
I really think it's not a good thing how professors seem to think something along the line of: "Oh, they've got holidays... Probably a good moment for them to be able to focus on their homework. Let's give them an extra long problem set!".
@BalarkaSen nope :P
Huy
Huy
@Danu isn't it locally quadratic or something like that?
We proved the general version.
Or some general version.
The thing I asked you about was a corollary.
I also gave you the general formulation we had.
It was that $\pi_1(X)$ has some universal property.
@Danu I agree.
@Danu Ok, fair enough.
Huy
Huy
I found interesting that the secant method converges locally of order of the golden ratio
19:54
That's precisely SVK in wrappers.
Mhm :)
But you can "see" the true version of SVK!!
It's very beautiful.
Siefert-van Kampen theorem.
19:57
@evinda No. $(ax_1+bx_2,cx_1+dx_2)$ doesn't denote an inner product, it's a pair of real numbers. Look at $\lVert T(x_1,x_2)\rVert_2^2 = \lvert ax_1 + bx_2\rvert^2 + \lvert cx_1 + dx_2\rvert^2$, and then apply Cauchy-Schwarz to each of the summands.
@BalarkaSen Seifert, you know, "e before i, except when it is the other way round".
@BalarkaSen Seifert-Van Kampen.
Oops, thanks @DanielFischer @Danu
Be careful :D
No, it's not Van Kampen.
Note the capitalization on the V, too.
19:59
It's van Kampen.
@Danu Is that NDR? It used to be lower case.
mmkay, found a nice example where Newton's method is problematic

« first day (1963 days earlier)      last day (3353 days later) »