Mathematics - 2018-09-24 (page 2 of 5)

1:00 AM

Having been an organizer several times, I speak from experience.

It's just sad.

Is Sir Michael Atiyah giving lecture on Monday Sept. 25 @ #HLF18? Yes. Will he presenenting a proof of the Riemann Hypothesis? Yes, that is what his abstract says.

Tweeted by HLForum on September 20, 2018 at 7:04 AM

Pretty sure they could have worked harder than to do this

Ted Shifrin

Chern still was doing mathematics until he died, but not making a spectacle of it.

Mike Miller

They're despicable people

Ted Shifrin

Organizers?

Mike Miller

specifically hlf. see above.

they're encouraging spectacle for their own gain.

Ted Shifrin

1:02 AM

Not sure how it's gain, but I don't know any details to discuss.

Mike Miller

whatever.

Ted Shifrin

There's a bunch more new people answering geometry questions, so I'm removing myself more :P

vzn

Atiyah riemann attack preprint(s) on google drive any opinion? reddit.com/r/math/comments/9icamx/…

manooooh

How many hours left in your country for Mr. Atiyah to make his speech?

Ted Shifrin

oy ... enough

Kasmir Khaan

1:07 AM

haha

Semiclassical

per one of the reddit comments: "How do we know this is actually him? A preprint on Google Drive rather than the arXiv? OK. Also this latex is weird (line skips between paragraphs), but I'm not in Atiyah's area so maybe that's just his style. Crackpot two-page "proofs" of RH are ten-a-penny. This doesn't even cite the papers he cites in his abstract."

I mean, if it turns out what he presents is identical to what's in the preprint, then that's proof it was him

manooooh

@Semiclassical he is 89 years old... we can not expect much from his ability to create content in $\LaTeX$

Semiclassical

Or it's not him.

I mean, maybe it is, but what exactly certifies its provenance?

Ted Shifrin

Even if I were to post a preprint to the arXiv, how would you know it's really me?

manooooh

What does it matter if he is or not?

Semiclassical

1:13 AM

uh

Ted Shifrin

Huh?

manooooh

We should judge the mathematical content and not the format

Ted Shifrin

This is a big tempest in a teapot, regardless.

manooooh

(I am not able to do it, but you surely yes)

vzn

agreed the provenance of the supposed papers/ google drive docs is not established yet. it would be nice if the links came from some blog, and the reddit post doesnt explain exactly where they came from, but presumably Atiyah is circulating it already.

Semiclassical

1:14 AM

Given that the story is "Michael Atiyah claims to have proven the Riemann hypothesis" and not "random google drive preprint claims to have proven the Riemann hypothesis"

it matters a lot

Ted Shifrin

Given that his last big claim was bogus, let's just let this drift ...

Semiclassical

"presumably"

manooooh

@Semiclassical he will be the author of the print...

Ted Shifrin

I think I need to cook dinner.

Joe Shmo

what was the last claim?

vzn

1:15 AM

@TedShifrin would like to see serious mathematicians dissect it. think that is the best response. their response might be short.

Semiclassical

Maybe. How do you know?

Ted Shifrin

@JoeShmo: Nonexistence of an integrable almost-complex structure on $S^6$.

This has been an outstanding problem for half a century.

vzn

@TedShifrin nobody has issued any written rebuttal so far on his sphere ideas. think the community should do better than that.

manooooh

@Semiclassical because of the tweet. Do not you trust serious institutions? I believe that they can modify the schedule, but the authorship is undeniable

Ted Shifrin

@vzn: The experts do not want to embarrass a legend.

Joe Shmo

1:16 AM

still is presumably?

Ted Shifrin

Yes, @JoeShmo.

Semiclassical

What tweet?

The reddit post didn't cite any tweet.

Kasmir Khaan

@TedShifrin Ted ! how can one prove that the space of functions with compact support , being a subspace of C_0 (R) , yet w.r.t same norm it is not a Banach space?

vzn

@TedShifrin then they are contributing to the confusion it would seem. aka put up or shut up™ in the vernacular...

Ted Shifrin

@vzn: I've actually talked with some of the experts, before you start to make a scene with no knowledge.

6

manooooh

1:17 AM

@Semiclassical they should... https://twitter.com/HLForum/status/1042670700652318720

Ted Shifrin

Well, when you're an expert, you can handle things the way you see fit.

vzn

@TedShifrin am not making any scene. there is really no scene except a lot of gossip. feel science proceeds through published papers. this happens in physics + cs why not math?

Kasmir Khaan

@vzn what does vernacular mean ?

Ted Shifrin

It certainly happens in math. And there is no published paper.

Eric Silva

i saw a robert bryant comment on why the sphere thing was wrong and i would trust him more on the matter than any 90 year old

Semiclassical

1:18 AM

@manooooh "Atiyah is presenting a proof of the Riemann hypothesis" != "This random google drive preprint is written by Atiyah"

vzn

@TedShifrin feel experts should seriously consider the scientific process. which involves peer review

Ted Shifrin

Yes, Eric, he's one of the experts, and he did try to be politic about it.

manooooh

@Semiclassical and why did you mention "This random google drive preprint is written by Atiyah"?

Ted Shifrin

@vzn: Fine. It was not published.

Semiclassical

Because that's what this conversation is about.

8 mins ago, by Semiclassical

per one of the reddit comments: "How do we know this is actually him? A preprint on Google Drive rather than the arXiv? OK. Also this latex is weird (line skips between paragraphs), but I'm not in Atiyah's area so maybe that's just his style. Crackpot two-page "proofs" of RH are ten-a-penny. This doesn't even cite the papers he cites in his abstract."

Eric Silva

1:18 AM

@TedShifrin i figured

vzn

how is publishing or not publishing a rebuttal either embarrassing or not for Atiyah? nobody can predict his feelings. seems like some kind of mass projection going on. isnt there embarrassment either way?

manooooh

Do not be impatient guys, we do not expect a big proof, but a simple one. His turn is coming

Joe Shmo

i have a question here -- Find derivatives of all order of $e^{z^2}$ at $0$. Taylor's expansion isn't particulalry insightful, $cosh, sinh, cos, sin$ seems like it could be fruitful, but im not getting anywhere

Ted Shifrin

@JoeShmo: Write out the Taylor series.

Joe Shmo

i did

Semiclassical

1:20 AM

I think that's not going to be helpful. The derivatives are related to Hermite polynomials iirc

Ted Shifrin

Huh?

Semiclassical

Oh, wait---evaluated at z=0

Ted Shifrin

Read, man.

Joe Shmo

yuh

playing around with it, without taylor, i get that the odd derivatives are $0$

vzn

← thinks peer review by hallway gossip is "embarrassing", "embarrassment" to science/ professionalism etc o_O

Ted Shifrin

1:22 AM

@Kasmir: Can you get a Cauchy sequence of functions of compact support (with respect to that norm) whose limit is not of compact support?

I actually don't see the answer.

Semiclassical

For clarification, this is what I had in mind: The Rodrigues formula for the Hermite polynomials is $$H_n(x) = (-1)^n e^{x^2}\frac{d^n}{dx^n} e^{-x^2}$$

Kasmir Khaan

first of all it is quite surprising that a subspace does not share properties with the big space :D

Joe Shmo

and the even ones seem to be coming from some combinatorial function (Hermite polynomials? some other stuff? I read about hypergeometric functions?)

Ted Shifrin

Fine, @vzn: Take your Ph.D. in mathematics and expertise and write an official evaluation.

The big space doesn't have a norm, @Kasmir.

vzn

@TedShifrin think your sarcasm is unprofessional.

Kasmir Khaan

1:23 AM

but ill try to think of something ! i see you are busy so we talk other time =p @TedShifrin

@TedShifrin it does, infinity norm

Semiclassical

And the nth Hermite polynomial is a polynomial of degree $n$. In particular, they're all either even or odd depending on degree

Ted Shifrin

No, @Kasmir.

Semiclassical

as such, $H_n(0)=0$ is definitely correct for odd $n$

Kasmir Khaan

||f||_ inf = Sup_ x in R |f(x)|

what is wrong with that norm Ted?

Ted Shifrin

Guys, we're doing derivatives just at $0$. Who cares about all this?

Because continuous functions on all of $\Bbb R$ needn't be bounded, @Kasmir.

Semiclassical

1:24 AM

Eh. I like it because it gives a context to why all the odd order derivatives vanish.

Ted Shifrin

That's silly, Semiclassic. An even function has all zero odd derivatives at $0$.

Semiclassical

derrrr

Kasmir Khaan

@TedShifrin but C_c ( R) is the space of continous functions with compact support not all the continous fns

Mike Miller

@TedShifrin can you just give anybody who talks about this for a week a half-hour tempban

Joe Shmo

for one thing, $1 = e^0 = e^{0^{2n}} = \sum_{n=0}^{\infty} \frac{0^{2n}}{n!} = \sum 0 = 0$

Mike Miller

1:25 AM

after a week the spectacle will have calmed down

Semiclassical

probably the real reason I like it is that physicists love Hermite polynomials

Kasmir Khaan

C_0( R) is the space of continous fns st f(x) --> 0 for x --> +- inf

Ted Shifrin

@MikeM: I won't be around to do that in consistent fashion, but it's tempting.

Kasmir Khaan

the infinity norm makes sense for me

Ted Shifrin

Oh, ok, @Kasmir. So they're bounded.

So that is a normed space.

Joe Shmo

1:27 AM

well except presunably $0^0 = 1$ in this instance

Ted Shifrin

Huh?

Joe Shmo

what in the world is going on

Semiclassical

@JoeShmo the zeroth term is just z^0=1

Kasmir Khaan

@TedShifrin Yes! anyway Ted i see you are super busy with others , ill continue digging on this one, help me when you are free later :D thanks again Ted <3

Joe Shmo

yes

Ted Shifrin

1:27 AM

I already said I don't see the answer, @Kasmir.

Joe Shmo

right @Semiclassical

Semiclassical

so it's really $1+\sum_{n=1}^\infty 0^{2n}/n!=1$

Kasmir Khaan

@TedShifrin Okay thanks anyway :D

Joe Shmo

yes

Kasmir Khaan

@Semiclassical Semi did you take fourier analysis ?

Semiclassical

1:28 AM

Not as its own class, no

Kasmir Khaan

being a physics guy am sure u did :D

Ted Shifrin

Are we really talking about what $\sum z^n/n!$ means?

Kasmir Khaan

oh okay ><

Semiclassical

But you kinda gotta learn it if you're doing physics, yeah

Kasmir Khaan

i heard that it is used for physics =p

Joe Shmo

1:29 AM

what do u mean

Kasmir Khaan

but what is you main subject?

you doing theoretical physics?

Semiclassical

well, tbh, it depends on what level of fourier analysis we're talking

physicists care about fourier analysis at the level of applications

Kasmir Khaan

am not sure how it is in USA

but here it is considerd first class in fourier , but you need to have alot of analysis before it

so atm am doing both analysis and fourier to catch up ><

Semiclassical

on the other hand, we typically don't take math courses in it, so our knowledge of how to do proofs in Fourier analysis is usually a good deal weaker

Kasmir Khaan

aha okay =p that is fine :D

vzn

1:31 AM

@MikeMiller ridiculous ban math chat in the math room

Kasmir Khaan

kasmir will continue his study ._. thanks for the talk !

Semiclassical

I think the point is that, within the first week, the discussion is liable to not be about actual math but rather just misinformed speculation and jumps to conclusions

Ted Shifrin

@vzn: You're not discussing math at all.

Semiclassical

I think the fuss about this 'preprint' is a sign of that, given that at this stage there's no particular reason to accept the provenance of such

Mike Miller

@TedShifrin he would have to start by knowing any

Ted Shifrin

1:33 AM

Indeed.

Semiclassical

(Though I am curious now if the preprint's statements regarding his ICM 2018 speech are correct. I mean, claims regarding 'understanding' the fine structure constant are usually taken as signs of crackpottery)

Ted Shifrin

@Kasmir: OK, i have a solution for you.

Kasmir Khaan

@TedShifrin Ted my hero :D pls dont tell me yet, i might find a way :D

Ted Shifrin

Here's a hint, @Kasmir. Start with a function in $C_0$ that does not have compact support. Try to get a sequence of functions with compact support converging to it.

@MikeM: I'm not fond of my new responsibility.

It was stressful enough being associate dept head for 8 years.

Cancer got me out of that. ;P

Mike Miller

Give it to @Daminark. What could go wrong?

Ted Shifrin

1:41 AM

LOL

Demonark has grown up a lot the last few years.

Daminark

Do I get power? :0

>:)

Ted Shifrin

I don't think anon or robjohn did that much officiating.

No, down, Demonark.

BTW, @MikeM, vzn left of his own accord.

But he was very annoying.

Daminark

Aw :/

Ted Shifrin

Work on your grad school essays, Demonark.

Daminark

Oh yeah I should probably start putting some time to that too

Mike Miller

1:45 AM

that guy has an opinion on everything

i wrote my research statement last night

get on my level daminark

Ted Shifrin

Cool, @MikeM. Feel free to send it to me for comments (probably won't have many).

Daminark

Brb getting on Mike's level

Ted Shifrin

shouldn't take too long, Demonark

Joe Shmo

hm @TedShifrin im getting $d^k/dz^k(e^{z^2})$ evaluated at $z=0$ is $2k \choose k$

sound right?

Ted Shifrin

First of all, it can't be right for odd $k$.

Mike Miller

1:47 AM

I'm on the third floor so you just have to take a couple flights of stairs

Joe Shmo

right.. for odd $k$ it's $0$

Mike Miller

@TedShifrin I'll FB you a copy.

Ted Shifrin

um, OK.

I think you're off by a factor of $k!$, @JoeShmo.

Joe Shmo

im getting $\frac{(2k)!}{(k!)^2} = {2k \choose k}$

Ted Shifrin

Why squared in the denominator?

I'll look after dinner, @MikeM.

Mike Miller

1:50 AM

Take your time.

I'm just playing video games in a hotel room.

Ted Shifrin

LOL ... I haven't been back there in almost 20 years. Sad.

There are some fabulous restaurants, though.

Well, there were. I'm sure there still are.

bolbteppa

@TedShifrin it's pretty sad to see these kinds of comments the past few days, this 'we don't want to embarrass a legend' thing, it's antithetical to science, whoever knows it's wrong and wont tell him should have higher standards, surely he'd be happy to be proven wrong

2

Joe Shmo

$f^{(k)}(z) = \sum_{n=0}^{\infty} \dfrac{(2(n+k))!}{(2n+k)!} \cdot \dfrac{z^{2n+k}}{(n+k)!}$

Eric Silva

i should get serious about grad apps now

uh oh

Ted Shifrin

@bolbteppa: People have told him. When people get older, they aren't necessarily in their right mind. It's a very touchy situation.

Eric Silva

1:53 AM

@Ted btw i think i will add upenn like u suggested

user525966

Is there something going on in MSE?

Ted Shifrin

OK, @EricSilva. Did you contact my friend?

Eric Silva

not yet but i was planning to

Ted Shifrin

Good, @Eric.

I haven't heard from him in ages.

@JoeShmo: The coefficient of $z^{2k}$ is what? You have two formulas for it. Don't give me formulas like the one above.

Eric Silva

oh whoa parsnips are good

Joe Shmo

1:54 AM

ugh i screwed it up

Ted Shifrin

Yup, parsnips are great ... especially mixed with mashed potatoes :P

Or roasted.

Eric Silva

my partner made a purée and i think i haven’t eaten these in like a decade

its delicious

Ted Shifrin

Ah, yup.

Mike Miller

i think we're going to a mongolian hot pot place tomorrow

Eric Silva

@MikeMiller omg gimme

Ted Shifrin

1:55 AM

Oh, that's new, compared to my time, @MikeM. Thai was the rage around '80.

I still remember a duck curry I used to get that was super spicy. Yum.

Mike Miller

That sounds delicious.

Ted Shifrin

@Eric: I've done parsnips or turnips in with potatoes for Thanksgiving.

<--- turns the room into food chat

Mike Miller

Better than the alternative.

Ted Shifrin

LOL

Eric Silva

the alternative being the status quo?

Ted Shifrin

1:58 AM

oh, mostly the math in here is fine

although sometimes it exhausts me

Eric Silva

@TedShifrin that sounds delish

i love me my roots

Ted Shifrin

if you move west, Eric, we'll have to collaborate regularly :P

Eric Silva

definitely

Ted Shifrin

Of course, SD isn't that close to Bay Area.

Eric Silva

i hope i move west tbh

closer to the bay area than chicago lol

Ted Shifrin

2:00 AM

well, get to work on your apps (and keep me in the loop)

Eric Silva

yeah i will keep you updated

user525966

this chat is making me cry lol

Ted Shifrin

Crying isn't productive

But I am leaving to cook dinner

Mike Miller

can you make a good bowl of borscht?

Ted Shifrin

Just a bowl? I've made borscht before.

Faust

2:02 AM

blarg

Ted Shifrin

I'm finishing my last bit of stuffed cabbage tonight. Same heritage.

Faust

Good morning @TedShifrin

Ted Shifrin

and howdy to you, @Faust.

You doing OK?

OK, @Eric.

Faust

yeah got my homework done

getting better at understanding things

Ted Shifrin

Still no manifolds homework?

Faust

2:03 AM

nope

Ted Shifrin

I'll have to send you mine, so watch out.

Mike Miller

@TedShifrin I can't find good borscht in LA. The last bowl I had was in the Polish airport and was truly awful.

Faust

but i did go through all the exercises in the sections we have covered though

that would be great

Mike Miller

They WAY overdid the dill. Made worse by the fact that I am not especially fond of dill to begin with

Ted Shifrin

Oh, I love dill, but typically borscht isn't dill-filled.

Joe Shmo

2:04 AM

i grew up in a ruski household

dill is the name of the game

Ted Shifrin

@JoeShmo, I'm 75% Russian, 25% Polish. But I cook mostly French, Italian, and Asian.

Did you fix that problem?

Joe Shmo

Russian and polish is the same kingdom at large, depending on where your great^n-parents came to america

$n\in \mathbb{N} \cup \{0\}$

Ted Shifrin

$n=1$ for me ... three from Russia, one from Poland, obviously.

Joe Shmo

stubborn latex

Mike Miller

@TedShifrin Certainly not when I want to eat it.

Joe Shmo

2:06 AM

nice

im first gen israeli

...gone american

Ted Shifrin

OK, @MikeM. I can make borscht if that's your first choice.

I'm about to leave. @JoeShmo: Did you finish that corrected computation?

Joe Shmo

not yet. staring at it

Faust

anyone heard about the supposed proof of the reimann hypothesis?

Ted Shifrin

The coefficient of $z^n$ is $f^{(n)}(0)/n!$, right?

Joe Shmo

because we're only interested in the $0^{th}$ term in that summation, right?

Ted Shifrin

2:08 AM

@Faust, we're done with that.

Joe Shmo

yes! ^

Faust

is it false?

heard the guys alittle senile

Ted Shifrin

So what is the coefficient of $z^{2k}$? Solve for the derivative.

We don't know, @Faust, but most likely ....

vzn

← very disappointed at elitism/ taunting/ sarcasm/ very cold edge etc in here, not sure what else to call it, outnumbered, surrender/ wave white flag for now :(

Faust

dam

was hoping itd be proven finnaly

Ted Shifrin

2:09 AM

@vzn: Elitism? Throw any more words around?

Joe Shmo

my capacity to brain had decayed in the past couple hours. ill have to pick it up again tomorrow

Ted Shifrin

OK, @JoeShmo ... I'm outta here now too.

Joe Shmo

night

Faust

d

peace ted

Mike Miller

@EricSilva i have kirby superstar

im going to that instead

i got rekt like halfway through starfox

vzn

2:11 AM

@TedShifrin lol and what age are you talking about? :P dont understand what exactly/ really makes it a "touchy situation".

Mike Miller

you have to beat the whole game in 3 lives which is 2 hard 4 me

CaptainAmerica16

2:31 AM

Hey everyone, I just started a room for anyone who wants to discuss the live stream of Sir Michael Atiyah's proof of the Riemann Hypothesis as it airs tomorrow, September 24th. I'm feeling a little skeptical. Come join the discussion if you're interested. I'm going to officially sign on about 20 or so minutes before the presentation begins.

Alex Clark

@CaptainAmerica16 Tomorrow = in 4.5 hours?

CaptainAmerica16

Depends where you live. Due my time zone means it'll go live at 3:45 am.

Alex Clark

Is that 4.5 hours away :P

CaptainAmerica16

Something like that, more like 5 hours.

bolbteppa

@TedShifrin having watched 5 minutes of his talk on the fine structure constant, I take it back :(

vzn

2:37 AM

Sir Michael Atiyah's Proof of the Rie…

For all those who are interested in Sir Michael Atiyah's proof...

CaptainAmerica16

What do you mean? Take what back?

Thanks for posting that. I'm new to chat rooms.

vzn

@CaptainAmerica16 np. another chat room is an excellent idea :)

Joe Shmo

2:50 AM

@TedShifrin so $e^{z^2} = 1 + \frac{z}{1!}\cdot z + \frac{z^2}{2!}\cdot z^2 + \frac{z^3}{3!}\cdot z^3 + \frac{z^4}{4!}\cdot z^4 + \ldots$

am I to understand from this Taylor's expansion (about 0) that $f^{(n)}(0) = z^n$?

$(f(z) := e^{z^2})$

this doesn't sound right, and in particular it fails to capture $f^{(n)}(0) = 0$ for odd $n$

Fargle

@JoeShmo The coefficients have to be constants for that trick to be valid, as far as I remember.

Joe Shmo

hm

Fargle

What you actually have is that $f^{(2n)}(0) = \frac{(2n)!}{n!}$, and that $f^{(2n+1)}(0) = 0$.

Joe Shmo

how did you get that?

Fargle

3:06 AM

Just by noticing that the coefficient of $z^{2n}$ is $\frac{1}{n!}$, and this should equal $\frac{f^{(2n)}(0)}{(2n)!}$ because it's a Taylor series.

Mike Miller

@Fargle Definitely, right? You would have to use the product rule.

Fargle

@MikeMiller Ah right. Thanks.

Mike Miller

It seems much better to use $e^{z^2} = \sum (z^2)^n/n! = \sum z^{2n}/n!$

Joe Shmo

yes, thats what im using

Mike Miller

Well, you're trying to think of $z^n$ as a coefficient, when it's not

The coefficients alternate between 0 and $1/n!$

Fargle

3:08 AM

In particular, $f^{(k)}(0)$ should never depend on $z$, because you're at the particular value $z = 0$.

Eric Silva

@MikeMiller i luv that game

Mike Miller

i did a little then swapped to contra 3

contra 3 is bullshit hard

Eric Silva

i’m bad at contra

actually i’m just bad at games, that’s why i like kirby, cuz its for babies like me

Mike Miller

i cant beat the first level lmao

Eric Silva

rekt

Joe Shmo

3:14 AM

@Fargle, @MikeMiller, got it.... the sequence is in some sense a spaced out version of $e^z$

Fargle

Indeed.

Joe Shmo

odd terms just don't appear

Rithaniel

3:25 AM

I tend to play more brain-hurting games, like TIS-100, myself.

Joe Shmo

Put another way, you can only associate the $z^{2n}th$ term to $f^{(2n)}(0)$. not the nonsense i was doing.

Mike Miller

@Rithaniel That game just felt like it was literally assembly programming

Spacechem was a lot more fun to me

Rithaniel

That because it was

Mike Miller

I wouldn't do assembly programming in my free time if you didn't make me pay for it

Rithaniel

Spacechem is awesome, too. Though, I've not beaten it.

Mike Miller

3:30 AM

I did long ago and then tried again and got stuck on world 6? or so

I dunno how younger me pulled that off

Rithaniel

My favorite game by Zachtronics is probably Opus Magnum, though.

(According to Steam I have 608 hours in that one)

I haven't really given Spacechem a serious try, though. I got to world 2 and then went back to world 1 and spent a week optimizing those levels.

Mike Miller

I seem to have 400 hours in fallout new vegas and 130 hours in Everyday Genius: SquareLogic, which is sudoku but sillier

very chill, sort of algorithmic like sudoku so it's not the braintwister you want

spacechem falls at 84h

Rithaniel

A silly sudoku sounds fun.

Mike Miller

Another puzzle game I love is "English Country Tune" by uhhh I forget his name

Pretty cheap and difficult and interesting

He also made a game called stephen's sausage roll which I never checked out but got good reviews (much more expensive, $30?)

oh cool I can buy ECT on my iPad for a dollar

im gonna do that

Rithaniel

Googling it, it looks interesting.

Zee

4:11 AM

@MikeMiller do you need good knowledge of algebraic geometry/complex geometry to go deep into symplictic topology ?

Mike Miller

Not really. They're different (but related) fields. On the other hand, deep in symplectic topology you tend to have to do (or black box) a lot of heavy duty analysis.