« first day (2635 days earlier)      last day (2393 days later) » 

7:02 PM
ok heres my work so far
am i on the right path?
$|\frac{x^2 + 5x}{x+1} - 3| = |\frac{x^2 + 5x}{x+1} - \frac{3x+3}{x+1}| = |\frac{x^2 + 2x - 3}{x+1}| = |x-1||\frac{x+3}{x+1}|$
we want to bound $|\frac{x+3}{x+1}|$
we have $|x+1| < 3$ and $|x + 3| < 5$
 
Looks good so far ...
Where did that come from?
 
oops i didnt add that we bounded $|x-1| < 1$
 
@TobiasKildetoft yes
 
anyways, am i right in saying we want to find a lower bound for $x+1$? since its in the denominator?
 
Hey everyone!
 
7:06 PM
Ah. So we're stipulating that $0<|x-1|<1$. YES, to that last one!
Hi Demonark. You purposely missed my point last night.
 
ok this is a lot less flurping confusing when i have no $\varepsilon$s
 
@Meow: I'm not just being arbitrary when I tell you that the paradigm should be followed.
Sometimes one learns and understands by following/imitating patterns.
 
ok so whats our lower bound for $|x+1|$
lets see
well $x$ is always positive from our $|x-1| < 1$ condition
so we can just find a lower bound for $x+1$, which is 1 as we just found
does that sound right?
 
Yeah. Can you write out a careful argument? 1 pt for basic inequalities; 2 pts for triangle inequality
 
umm
let me get the notebook out
applying triangle inequality to $|x+1|$ doesnt look promising since then we get an upper bound
 
7:15 PM
Hi, can you suggest some books for this ?
Also, a nice joke that I learnt today:
 
i dont know of a better wya to phrase it
 
@Meow: The reverse triangle inequality is super important, shows up all over multivariable, too :)
 
You need balls to do combinatorics.
 
i dont know of a better wya to phrase it
ok let me try that
$|x+1| \geq ||x| - 1|$, since $x$ is always positive $|x| = x$ and $|x+1| \geq |x-1|$ so $|x+1| \geq 1$? something about that seems fishy
 
No, always write the thing you're doing in terms of the thing you know, @Meow.
$x+a = (x-1) + (a+1)$.
 
7:20 PM
o/
 
$|(x-1) + 2| \geq ||x-1| - 2|$ somethng like that?
 
Heya, @Danu. Enjoying what's left of your last vacation ever? :P
 
You're horrible :P
 
Well, that's why you endure me.
Yes, @Meow.
 
@TedShifrin Thanks for the answer ! BTW, I have a book of almost the same name (40 Challenging problems in probability) from Springer too ! Does the book by Pitman addresses the issues commonly faced be beginning learners (and also understandable by beginners) ?
 
7:23 PM
I don't know the Springer one, @Alex. The texts I named are written for college students, but they don't use a huge amount of baggage (although eventually they get to using calculus). I think Pitman has lots of interesting stuff in it; he's more of a statistician, so it's a bit more concrete than some math texts.
@Meow: So do you see the end?
@Alex: A few years ago I decided to teach probability for the one and only time before I retired. I truly enjoyed it and wished I had done so many years ago. It can be confusing, but it's very important (things like false positives/negatives on tests) and interesting.
 
It wasn't really on purpose, I'm still not sure what you were getting at
 
well i know the inside has to be negative but im not sure thats what youre looking for
 
So what's the absolute value, @Meow? You want a lower bound on the original thing, remember?
Demonark: Look in the star-list.
 
Oh stacks
 
mutters annoyedly about Demonark's density
 
7:28 PM
Ahah, that was a good one
Give me some slack, I'm on my phone so I have to actually look for the starboard
 
Hey
 
I learned long ago never to give you slack, Demonark.
 
So when I just saw that I didn't make the connection. Also yo
 
Hi, Balarka.
 
Hey, I'm reading a calc of variations intro pdf but something it says confuses me
It says that the function inside the functional isn't a functional because the range isn't $\mathcal R$ but.. Why isn't it $\mathcal R$?
 
7:30 PM
Usually functional is reserved for the function on the space of functions.
(Like the thing you're trying to optimize.)
 
I get that, but the pdf doesn't make a statement about the domain being wrong, just the range
 
@TedShifrin Ah very nice ! BTW, do you know similar books for linear algebra (because I need linear algebra to understand the very nice stuff here math.cmu.edu/~ploh/docs/math/mop2011/alg-comb.pdf , but they're way over my head)
 
Do they mean to say that it's not $\mathcal R$ for all functions?
 
Without reading the stuff in context, @Phase, I can't answer your questions.
 
Ah sorry
Page 4, the footnote
 
7:31 PM
I mean that's probably fair but... :'(
 
@Alex: Of course my favorite linear algebra book is my own :P
 
oh heh you wrote one !
Duh duh way over my head. I surely can't understand what you wrote in the book as.wiley.com/WileyCDA/WileyTitle/productCd-047152638X.html
 
Not that book, @AlexKChen. The linear algebra one.
@Phase: Looks like crap to me.
 
Oh
:(
 
@TedShifrin That one right ? The "manifold" in the title is telling that it's not exactly for beginners.
 
7:35 PM
I just answered you on that, Alex. It says specifically Linear Algebra as the main title
 
Hi chat
Hey @TedShifrin
 
@Alex: It's a matter of taste. My coauthor and I emphasize dot products and vector geometry from the beginning, whereas a lot of linear algebra texts delay dot product until Chapter 8 or later ...
heya @Lozansky.
 
So it turns out that in order to try to make supersymmetry rigorous, one needs schemes
 
@Phase: The issue is not the range but that the domain should be a space of functions, rather than a finite-dimensional $\Bbb R^n$ (in that case, $n=3$).
 
I gotta learn some commutative algebra now
 
7:37 PM
Guess so.
 
how do I Spec
 
You look at prime (or maximal) ideals of your commutative ring.
 
sorry im back
 
I'm gonna leave soon, @Meow. It's past my lunchtime. Finish up my last question :)
 
@Danu Rekt
 
7:38 PM
okay
 
In case you didn't find it, Meow: So what's the absolute value, @Meow? You want a lower bound on the original thing, remember?
 
Rehi chat
 
Maybe Danu and Balarka can learn schemes together.
off running over innocent people, @Alessandro? Oh, no, way too late for that.
 
Nah, I was busy drinking mulled wine, I'd better not drive after that
 
I wish I knew more algebra
I wish I knew more analysis too
 
7:40 PM
@TedShifrin but it defines the domain of $F$ as {y(x) : y:[a,b] -> R}
 
It's not cold enough here for that.
 
Is that not sufficient?
 
@TedShifrin I'd first need to go through basic algebra
 
Is it obvious that $\cos(x)$ is a solution to $y''+ay'+by=0$ if you know $\sin(x)$ is a solution?
 
@Danu: There's a lot of "basic algebra" that really doesn't get used much in commutative algebra.
 
7:40 PM
For $a,b \in \mathbb{R}$
 
Yeah, by complex conjugation, @Lozansky, most likely.
 
@TedShifrin True... Should I just start with Atiyah & MacDonald?
 
@Lozansky Sure, sin(x) = cos(pi/2 - x). It's a solution with a different initial condition, I believe
 
I guess it's something something eigenvalue of $i$ so $e^{ix}$ is a solution and so the real and imaginary part form a set of solutions
 
It's terse. I sorta like Eisenbud's book, which has lots more examples and words, but it's the opposite of terse.
 
7:42 PM
$||x-1|-2| \geq |1-2| = 1$
 
Eisenbud is a big big big big tome
 
Hmm, @Meow. How did you get the $\ge$?
 
I like it, but A-M is concise, eg
On the other hand hard to read
 
@TedShifrin The "with a view towards AlgGeom"?
 
Yeah.
 
7:42 PM
most of the examples and exercises are refereed to exercises, etc
 
Hmm okay, guess it's kind of obvious
 
I honestly think Eisenbud is more a reference book than a textbook @Ted
 
nevermind its $>$
 
@Danu: I sorta wish I hadn't given it away.
 
That book is f*cking huge doe
 
7:43 PM
because its $|x-1| < 1$
 
Can't just casually read 650 pages
 
I still am not sure how you justify the inequality.
@Danu: A-M is tiny, but you can't sit down and just casually "read."
 
@Danu I can suggest you something equivalent to A-M but with lots of geometric intuition and classical stuff.
 
Yeah
I see
 
Reid's commutative algebra
 
7:44 PM
You need to work out every sentence and do 20 exercises to get things.
 
That's not too bad @Balarka
 
We're more or less following Reid's too in the commutative algebra course
 
Sounds like analytical mechanics...
 
It's full of beautiful pictures. I think if you read Reid, you will not miss out on any important commutative algebra that's in A-M.
 
Oh, sounds like Alessandro should tutor Danu :P
 
7:45 PM
@TedShifrin The algebra force is strong in this chat
 
Rather weak, in my case.
 
@Meow: I guess I'm saying I would write $||x-1|-2| = 2-|x-1|$ and then it's clear.
 
its like
 
Danu's gonna make me miss our complex geometry days.
 
anyways continuing with that proof earlier
we have the lower bound of $|x+1| > 1$ so
$\frac{1}{x+1} < 1$
so $|x-1|\frac{|x+3|}{|x+1|} < |x-1|\frac{5}{|x+1|} < 5|x-1|$?
does that look right?
oops wait DONT YELL YET
 
7:48 PM
@TedShifrin Hey don't worry
My actual PhD project is pretty close to what's called special complex geometry
 
That looks good, Meow.
 
Complex geometry with a distinguished connection such that $d J=0$, where we view $J$ as a $TM$-valued one-form
 
So $d$ means $\nabla$?
 
$d_\nabla$, not $\nabla$
 
I guess they usually write $d^\nabla$?
 
7:49 PM
exterior derivative one bundle-valued forms
 
Right.
 
ok so we want this also to be less than $\epsilon$ so we can complete our inequality
 
Right, Meow.
 
I guess it's the same thing, depending on notation
 
so $|x-1|$ also $< \frac{\epsilon}{5}$
 
7:50 PM
And your final beautiful sentence is ...
 
[perfect setup]
 
So, @Danu, are you excited about the new chapter in life?
 
Mostly sad to be leaving, honestly.
 
Aw.
Is your girlfriend staying in München?
Hmm, guess I shouldn't have asked.
 
yup
 
7:53 PM
It's hard to get to a state of permanence in academia (I suppose in most jobs) ...
 
Hm, I have to decide what to do now
 
wonders if Meow is writing his beautiful sentence or if I should just go eat lunch
 
Choose $\delta = \min(1, \frac{\epsilon}{5})$. We have that $|\frac{x^2+5x}{x+1} - 3| = |\frac{x^2 + 5x}{x+1} - \frac{3x+3}{x+1}| = |\frac{x^2+2x-3}{x+1}| = |\frac{(x+3)(x-1)}{x+1}| = |x-1||\frac{x+3}{x+1}| < |x-1|\frac{5}{|x+1} < 5|x-1| < 5\cdot\frac{e}{5} = \epsilon$
ow my eyes
 
I could read some Forster for a while, do some fiddling with Galois theory, try to type my ODE notes, or listen to completely hipster music
 
BEAUTIFUUUUUL
 
7:54 PM
@Balarka: I thought this was a cool question.
 
how does that look :)
 
You could break it into two lines, Meow. All you're missing is: If $0<|x-1|<\delta$, then we have ....
Otherwise perfect (except for typos).
 
nice, ill do the rest of the rational things after my nap
 
@TedShifrin Having $k_1 = 1$ and $k_2 = -1$ at every point would mean $K = -1$ everywhere, would it not? But there's no compact everywhere nonpositively curved dude out there in R^3
 
@BalarkaSen hipster music for sure :P
 
7:56 PM
Do that big balls argument
 
@Balarka: He never said compact or complete.
He already knows that argument.
 
@Meow: Good progress. Nap well :)
 
thanks!
 
@TedShifrin I actually don't know how to prove there's no complete everywhere nonpositively curved surface in $\Bbb R^3$, I don't think. But I see what he's asking; I know everywhere -1 curved surfaces in $\Bbb R^3$ but none that I can construct off the top of my head with $k_1 = 1$ and $k_2 = -1$, nope.
@Daminark I have to look for some newer hipster music.
 
8:01 PM
Of course there are complete surfaces of nonconstant negative curvature.
 
Maybe some psychedelic death metal or some shit
 
Ugh.
@Balarka: Hilbert's theorem is only about complete constant negative curvature.
 
@TedShifrin Ahhhh
The last two statements are clarifying
 
Hyperboloids of one sheet, saddles, etc., give you plenty of examples.
 
Right, I was a little confused for a while.
 
8:03 PM
Without my classification exercises, Codazzi should tell you enough to see it has to have curvature 0, which is oops.
I am, however, extremely fond of those exercises :P
OK, lunchtime ...
 
I should work them out at some point
Enjoy!
 
BBIAB.
 
Gonna give it another shot - anyone here familiar with maximum likelihood estimation?
 
8:18 PM
@Studentmath Depends how advanced the question is. Hahaha
 
@ÍgjøgnumMeg I'd say pretty simple. But I just don't get the basics so everything is hard :)
To the point, Assume I have $X_1,...,X_n$ from $N(0,1)$.
I define $Y_i=0$ when $X_i\le 0$ and $Y_i=1$ when $X_i>0$.
$\theta=P(Y_1=1)$
I want to do an MLE on $\theta$. I am really clueless. The probability guy in me screams this is all wrong.
I just don't see which function should I build or can I build..
 
Could someone explain to me / refer me to something explaining the basics of, and consequences of, Cauchy sequences in a space?
 
@Studentmath Just as an observation... your $\theta$ is equivalent to $P(X_1 > 0)$ and $X_1 \sim N(0, 1)$
@Studentmath Unless you meant $\theta = P(Y_i = 1)$?
 
@ÍgjøgnumMeg I don't think so. I think they do mean it.
But then again - if that's the case, I don't get why to make an MLE about it.
It's easy probability. Perhaps it is a mistake, but then how can you even explain the expression if $i$ is not fixed...
 
@Studentmath Me neither unfortunately, your $\theta$ doesn't seem to depend on anything except the value of $X_1$, but you're given the distribution of $X_1$
 
8:31 PM
Exactly..
 
8:42 PM
Perhaps I should post it on the forums, maybe someone will point to something I'm missing
Oh god I am so stupid. It's $N(\psi,1)$. $\psi$ is unknown.
Now it makes sense...
 
@Studentmath Hahaha, okay that's better
 
@ÍgjøgnumMeg Yeah ;) Though I am still unsure of how to build the estimator.
 
8:57 PM
hi chat
 
I asked a while ago, but how can I render mathjax in the chat rooms? I'm using chrome on Android. The solution (tinyurl.com/cfqcvpc) in the description doesn't work for me.
 
@MatheiBoulomenos Hey :)
 
Hi @MatheiBoulomenos
 
@MatheiBoulomenos I have a question for you!
 
Sure, I love questions
(at least if they're about algebra or number theory :P)
 
9:02 PM
@MatheiBoulomenos It's a soft question unfortunately ;( In Germany, when you say "take a seminar" what do you mean? I feel like the word is used differently over there. Is it just like a reading course?
 
No, it's not a reading course
Everyone who attends a seminar gives at least one talk
the others listen and ask questions
mostly listen
 
I see; the structure of the BMS programme requires you to take at least one seminar I think, so that just involves giving talks and listening to talks?
 
yes, at my uni you have two do two seminars in your bachelor and two seminars in your master at my uni
a lot of people I know (myself included) do more
But yes, it's giving talks and listening to talks
 
Ah okay. I only have to give one talk for my entire degree and it's just a presentation on the dissertation. hahaha
 
Hey @Mathei!
 
9:07 PM
Well, you have to present (or as we say "defend") your bachelor and master theses as well
Hey @Daminark
 
Yeah we call it a defence too, but it's not really a defence because the lecturers don't really care. You just show you did some work and pass. hahahah
 
How's everything going?
 
kchaoungkhpha Top of the morning to you laddies
 
Great! I love algebraic number theory and I'm really glad that I can finally take a lecture on it. I'm a little nervous about my seminar talk this tuesday, though
And how's it going for you? @Daminark
 
Pretty well! Complex is fun, algebra is starting to get better, I hope.
And topology is sick
 
9:17 PM
pls tech me algbr
 
Lol, we only barely got to like, homomorphisms
So we were phrasing group actions this whole time without that
 
Group actions before homomorphisms? interesting
 
That is indeed a curious approach to things
 
Yeah, he just spelled out the map $G\times X\to X$
 
I sympathize with the approach
 
9:20 PM
Right.
 
blorp
 
Hi @Mike
I end up reading Forster it seems
back to dealing with a million tangent spaces
 
@Mike I'm not sure what I think of it, though I'm finding that the class in general just seems to going poorly. Between the pace, the problems, all that business, I'm finding that a number of people are just starting to really dislike the subject
 
dang
 
I think it's going super-slow, above all the problems, right?
 
9:26 PM
Some people are just predisposed to dislike it but given that the rate is much higher now than what I think the class last year had, I do think part of it is just that the presentation is iffy
 
Is Rolle's theorem a special case of The Mean value theorem ?
 
Yes.
 
"Rolle" is the right spelling, iirc
 
@Semiclassical Let's Rolle with that
3
 
So, part of it is that, the other is that the problems have often been unfun. He's given us some really good ones for what it's worth
 
9:27 PM
Haha @BalarkaSen
 
Nobody should be made to dislike algebra. That's a crime :'(
 
But it Rolle's theorem adds the condition $f(a)=f(b)$ which isn't required in the "general" case ?
 
@Daminark Gotcha
 
This most recent pset had Burnside's Lemma, one about relating a certain action of S_n to matrices (quite long to state so I'll do it when I'm not on my phone), a few that used Burnside, etc
 
my algebra course was great, the only thing I disliked was that the exercises were too easy
 
9:31 PM
But I've found many people didn't like the problems about rotation groups, and some were just like "List out the number of elements in each conjugacy class of S_6" which were long and dull
 
My algebra course was awful and the course I'm due to start next semester will be painfully easy too. :(
 
The pset this week is extremely short, though it did have a problem I quite liked
 
@Daminark Oh yeah, that sounds dull. We had to classify all subgroups of S_4 at one point, which involved a lot of casework, but it's useful for computing Galois groups of degree 4 polynomials to have that result, so there's that
 
But yeah we've had one problem on our midterm which was just horrendous
It was like, list the cycle types in S_{26} whose elements have order 75, 76, 77, and 78
Hopefully the class will get better soon though
 
that's not even algebra, that's just combinatorics
 
9:37 PM
To be fair that's half of introductory group theory
 
Not the way I used to teach it.
One needn't spend much time on the symmetric groups, in fact.
 
You can use group theory for combinatorics, but you don't have to
 
Matrix groups are more important to many of us.
 
Yeah I guess unless you do A Geometric Approach$^{TM}$
 
We didn't even mention Burnside's lemma
 
9:38 PM
Matrix groups have almost never come up with us actually
 
Burnside's Theorem is so cool, though, Mathei.
 
I made a TM joke involving the tangent bundle once when explaining it to someone
Very satisfying
 
Like on the second day of class we were just listing some examples and GL_n came up
But we've mostly been working with S_n, D_{2n}, cyclic groups, and rotation groups of the cube/tetrahedron/dodecahedron
 
I know it's cool, but the course was already packed with cool stuff, so I guess it didn't hurt to leave it out @TedShifrin
 
And really mostly the cube
 
9:40 PM
I only mentioned it because of combinatorics.
For my course it fit naturally because of all the symmetry groups of objects we'd done.
 
We worked with dihedral groups, cylcic groups, symmetric, the quaternion group
 
But the class has been a lot of counting, like we spent a whole day once finding how many elements there are in a given cycle type
 
That should be homework, Demonark.
Bleh.
Certainly doesn't sound like Herstein. What book?
 
Dummit and Foote
 
That's not the flavor of D&F at all. Weird prof.
 
9:42 PM
We didn't do much symmetry objects, we did not have a geometric approach$^{TM}$
 
Well, your loss, Mathei.
 
Herstein isn't what we use here, it's just what I first learned it from when I was in the REU
 
Herstein was at UC, of course, so I wondered how long they'd use it. I am not so fond of his book, but it has some great exercises.
 
But yeah our prof just did stuff on his own more than follow the book. I don't even know what quaternion groups are
 
9:43 PM
Like Balarka, I keep going back to Artin, because he shows you algebra as mathematics, not as technical formal things.
 
I find it really strange that most people seem to teach out of a book. In my uni people usually do their own thing
 
I love Artin; it ties a lot of math in togather
 
But yeah we have our first 3 psets up on this website so you can decide if you like the problems or not. I felt they were on and off
 
Mathei: I did that and wrote my own books for four of the courses :P
 
9:45 PM
@TedShifrin sure, but you can do it without writing a book, which is what my profs do, mostly
 
For certain courses, I couldn't improve on the books much — Munkres, Guillemin & Pollack. Graduate complex I was never happy. I didn't love Ahlfors or Stein/Stakarchi.
 
As for teaching out of a book, I've seen a lot of different styles
 
Even when I wrote the book, I never just read out of the book to the class, but I know some profs who literally put the book up on the board every time.
 
My analysis professor taught and assigned all his problems from these 4 books, and aside from a few tangents even the theorem order was mostly the same
Another was like "I'm just kinda doing stuff, this material is all in finitely many books so I can give you that"
@TedShifrin I don't think I'm terribly fond of that style
 
I know I am not, Demonark.
 
9:50 PM
@TedShifrin I'm sorry for bringing this up (again) but I still haven't understood why is that the number of rational numbers above the line $y=\epsilon$ being finite matters, I talking about that proof we discussed perhaps 2 days ago, thanks in advance.
 
It's driving me nuts teaching off slides prepared by someone else, too. A whole new experience.
 
that's just lazy
 
Oh, no problem, @Fuzzy. So let's say that $x_1,\dots,x_s$ are those finitely many $x$'s with the dots above the line.
@Mathei: I have no option in what I'm currently doing.
 
We're trying to show continuity at $a\notin\Bbb Q$.
 
9:51 PM
Wait slides? Not even a chalkboard?
 
@Mathei: I'm doing a high-school-ish course that they want to be universal no matter who teaches it. But of course the discussion and what we write down on each slide is up to me, more or less. But still ... A white board for little discussions, Demonark.
@Fuzzy: Take $\delta = \min(|a-x_1|,\dots,|a-x_s|)$. What can you say about $f(x)$ if $|x-a|<\delta$?
@Demonark @Mathei: I am writing a few extra problems that are my favorites or ones I'm thinking up that are a bit different, but interesting.
@Demonark: I am getting better at writing on computer screens with pens that show up on the wall :P
 
God I much prefer stuff like, what I've seen from you on YouTube, as well as Laci and Schlag
 
No chalkboards is strange
 
Me too, Demonark, but I knew it was a non-traditional teaching set-up when I took the job.
 
Well if $|x-a|$ is smaller than all of the $x$'s mapping to the dots above $y=\epsilon$, then $|f(x)|$ must be smaller than $\epsilon$. @TedShifrin
 
9:57 PM
And the little kidlets are pretty good. Although the best one missed class last week, and I'm worried it's too easy for him and he's disappearing.
@Fuzzy, certainly when $x\in \Bbb Q$, and if not, even easier. So that proves continuity at $a$.
 
More participation fueled, not too rigid. And that's unfortunate, is it that the content is just not what he was looking for?
 
He seems to know an awful lot, Demonark. He knows esoteric stuff I've never thought about, too. And the few times I've asked him to derive/explain, he does it, and well. When he writes exercises in class that ask for explanations, he writes full sentences. Pretty impressive 9th grader.
I've been trying to give him some harder stuff to challenge him. We'll see if he does this week's homework before tomorrow morning.
 
Damn, nice
 
We're getting to harder stuff now ... trig addition formulas and applications. The proof we'll do in class together is one I've never seen (although I'm kicking myself for not having thought of it).
 

« first day (2635 days earlier)      last day (2393 days later) »