The h Bar

YouTube Rewind: The Shape of 2017 | #YouTubeRewind

3:26 AM

OMG! The attack of the spam bots has begun.

...***over 10,000*** on SuperUser...

in The Clubhouse, 5 hours ago, by Philip Kendall

@RamChandraGiri Right, I deleted the most recent 50 or so spam accounts but after that even my trigger finger got bored. As noted, this is network-wide (SuperUser has had something over 10,000 of this accounts created!) but not causing direct harm as all they're doing is using their profile to link.

What's going to happen to Winter Bash if they take over?

:P

user228700

4:50 AM

Eyy, YouTube Rewind is out!

user228700

►

> no jake paul on the cover

trash!

user228700

Regardless of that, it's not very good.

hmm

user228700

Gosh, they keep getting a little worse every year :-/

5:01 AM

is there a room for Machine learning or data science?

chat room

@Kaumudi.H I don't even know what this is

user228700

Huh?

Machine Learning

General discussion for machinelearning.stackexchange.com

@Kaumudi.H youtube rewind

it's after my time

user228700

Lol.

5:06 AM

Morning all

Data Science SE

For general discussions about the site and data science in gen...

yo

greato

last message 122 days ago...

uh...

@JohnRennie any idea how to get TeX to separately label the components of a piecewise equation?

machinelearning.stackexchange.com

5:09 AM

@0celo7 TeX is largely mysterious to me I'm afraid

user228700

@JohnRennie Morning! :-)

@Kaumudi.H Refreshed after a good eight hours sleep? :-)

user228700

Sort of :-P

21 Things that Turned 21 This Year (2017) - mental_floss List Show Ep. 523

7.5? :P

user228700

For those older than 21 (to be able to remember these):

user228700

5:11 AM

►

user228700

@user685252 :-P Yes.

:-D

7.4999...

area51.stackexchange.com

5:27 AM

uhhhh

I'll try the data science chat

I have a bunch of questions that I don't think would do great in terms of creating a question in an actual stack exchange lol

askaway

You're familiar with LSTMs?

nope, sorry

oh

I thought you meant for me to ask you XD

I did :-)

5:34 AM

Storm Caroline hits UK!

huh...

en.wikipedia.org/wiki/Long_short-term_memory?wprov=sfti1

:: John prepares for Storm Caroline by yawning, making another coffee and scratching his bum ::

Well my question is, in a LSTM or even simple RNN model implemented on something like Tensorflow or Pytorch, you are asked to break up the sequence into sub-sequences and feed them into the LSTM cell in batches like : [batch_size,sequence_size,sequence_element_size] and then the LSTM is run simultaneously for each sequence in the batch from left to right. The LSTM does this by keeping batch_size number of hidden states and cell states and feeds these from one time step to the next.

However, this seems to suggest that information is not fed from the end of the first batch to the beginning of the next batch

The snow could make it especially nasty @JohnRennie

5:38 AM

backprop is also truncated

Isn't that suboptimal?

I don't think this question is answered on the wiki page for LSTMs

since it deals with implementations of LSTMs in a specific language

5:52 AM

@JohnRennie ::Mithrandir prepares by flying off to the tropics::

@Mithrandir24601 quick

Lagrangian variable moves with the flow?

or is that Eulerian

@Mithrandir24601 you certainly picked the right time to go :-)

Euler

@JohnRennie It's going to be so warm :D

:: John wonders what the customs will make of the small bag of icing sugar he has slipped into Mithrandir's luggage ::

6:02 AM

@0celo7 It's 6am -my bios is still starting up, so I'm running a preprogrammed program from ROM :P

@JohnRennie they're going to have a field day on the way back with all the tea I'm going to buy :)

Is it a business (i.e. research) or pleasure trip? You did say something about it earlier (Japan? China?) but I've forgotten the details ...

@user685252 keep nuking them.

@Sid :)

@JohnRennie conference in China - I'm arriving a day early to, erm... 'rest after the flight' :P

And then staying on a couple of days afterwards :)

@Mithrandir24601 I'm seriously envious. The farthest East I ever got was Thailand.

6:13 AM

@JohnRennie this will be the farthest from home I've ever been

@JohnRennie Thailand is almost at the same longitude as China

@Sid Farthest East in a cultural rather than geographical sense :-)

@Mithrandir24601 better start smoking

the filter will protect your lungs from the smog

Hmm, I should probably get out to visit other countries soon...

@0celo7 that's my one worry. Especially considering that I actually have a variant of bronchitis :/

6:18 AM

One worry?

Hm, now that I think of it, one of the worse things about the East is our high levels of pollution

How about getting your identity stolen the moment you turn on your devices?

China has state-sponsored espionage programs targeting foreigners

@0celo7 That shouldn't be a worry as long as you don't know that they are spying on you.

Ignorance is bliss.

@Mithrandir24601 they will find your search history

@0celo7 not bringing my laptop and it's not like there's anything to worry about on there anyway

6:25 AM

@0celo7 I think you might be mixing up China and North Korea :-)

@JohnRennie no

don't defend China

no, instead defend trump :P

if the argument is good, he can do that

I am pretty sure China spies on you iff you have a known history of criticizing their Govt.

iff?

why

surely they try to spy on everyone

our government certainly does

6:29 AM

Pretty sure, the Chinese Govt. has better things to worry about than spying on a nobody like Mithrandir246.

@Sid that's what people said about the NSA

My tin foil hat is firmly in place

lol

Morning Mr.@JohnRennie :-)

perhaps, the NSA is behind the 10,000+ bots on SuperUser?

@Alan Morning :-)

6:32 AM

who knows

@0celo7 USA has like a complete law which legalizes such spying...

@JohnRennie :D

May I ask another question? O;)

SHM again :3

Of course, yes.

"Lissajous figures"

@Sid You think China has laws against government spying? Are you kidding me?

6:33 AM

@Alan Yes ... ?

@0celo7 No. But, I don't think they have a law which legalizes spying either. It's like a grey area. And everyone knows on which side the Govt. leans

@Sid Look, none of the great powers have laws against spying on other countries. Why do you think the US govt. can legally drone people in Pakistan as long as they're not citizens?

You need a Supreme Court to decide that.

@JohnRennie For the composition of SHMs in two perpendicular directions, x=Asin(wt) and y=Bsin(wt+phi), the resultant Lissajous figure is some sort of an ellipse, right?

6:36 AM

People from other countries are not people (whether that's good or not is debatable)

THERE IS NO SYSTEM BUT GNU AND LINUX IS ONE OF ITS KERNELS

@JohnRennie Morning John :)

@BernardoMeurer Hail Richard Stallman

@BernardoMeurer Morning :-)

@0celo7 because the US can get away with almost anything.

@Sid and so can China or Russia

6:37 AM

@Alan yes

Russia invaded Ukraine

No one did anything

@JohnRennie If I go to the UK will you take me to Greggs and feed me?

If you visit the UK we can do a lot better than dining at Greggs :-)

I don't want to dine there, I want breakfast :P

@0celo7 I guess so.

6:38 AM

I want a picture of you and me in adidas pants crouching near a tree eating something from greggs in the cold

I will frame it and cherish it

Russia is different.

I think I had in mind a stack of bacon, eggs and pancakes so high I am barely visible behind it ...

I lost my folder with all of my analysis and topology notes

this is a disaster

@JohnRennie Deal

(contd.) And if there were no phase difference between $x$ and $y$, the Lissajous curve would've been a straight line [I'm getting y = (B/A)x ]... I'm told that in this case, I'd be dealing with Oscillatory motion, but not in the general case (ellipse).

6:40 AM

@0celo7 What

For real?

Don't you type them up?

Question: Something very obvious must be evading me... I don't see anything that looks like an SHM when the phase difference is zero :/

@skullpatrol I don't have dozens of hours to waste

@Alan Do you have access to a spreadsheet app like Excel, or even Google Sheets? If so I strongly encourage you to have a go at plotting these figures to see what they look like.

Ah? How exactly do I "plot" them? (Pardon my ineptitude, still a school student :-) )

6:42 AM

You are quite correct that with a phase difference of zero you get a straight line on an xy graph.

You're obviously a fast enough typer in this room :P

@JohnRennie :-)

But remember that both x and y are functions of time.

where the heck could they be

If you calculate x and y as a function of increasing time you find the point (x,y) moves along the straight line and back with time i.e. it oscillates between the two ends of the line.

6:44 AM

Yes, and I substituted functions of (wt) with other functions (in terms of displacement)

@JohnRennie <Need to think about it>

A minute :-)

Ah yes, it oscillates... but is it SHM (doesn't look like it)?

A hint would do ;-)

Retrace your path from the last place you saw them @0celo7

@Alan Let's call the distance from the origin $r$. Then $$r^2 = x^2 + y^2$$ Yes?

@skullpatrol a year ago in a different house

pretty hard

:-/

@0celo7 okay, you are doomed

6:47 AM

@JohnRennie Yep, I'm good :-)

oh fuck

@BernardoMeurer rebecca has them

@Alan And $x=A\sin(\omega t)$ and $y=B\sin(\omega t)$. Still with me?

@BernardoMeurer nvm they were underneath my desk

If we substitute in our equation for $r^2$ we get: $$ r^2 = A^2 \sin^2(\omega t) + B^2 \sin^2(\omega t) = (A^2 + B^2) \sin^2(\omega t) $$

6:50 AM

Ah wait, I think I got it :O

@JohnRennie Yup :D

and $r$ is a function of sin(wt)

Hence an SHM?

@Alan I was wondering when you'd jump ahead to the solution :-)

Yeah, I'm kinda slow :3

Thanks yet again :-)

it's nice when the phases line up

sips coffee

Coffee, that's a good idea. Better have more coffee to fortify myself for the storm ...

"the storm"? :D

6:53 AM

I knew you would find them once you put your mind to it @0celo7 :-)

Well, I'll be off then! Thanks Mr.@JohnRennie :-)

@Alan *Dr.

Esquire :P

Dr.@JohnRennie it is then! :O

John Rennie Esquire MA PhD Cantab

6:58 AM

@JohnRennie wait you're an Esquire??

@0celo7 It just means Mr in the UK

i.e. Mr John Rennie = John Rennie Esquire

oh

Esquire means respect.

not here

@JohnRennie what is Cantab?

6:59 AM

here it means "slime ball"

John Rennie Slime Ball MA PhD Cantab

better

@Sid Cambridge University

@0celo7 ah, Wikipedia tells me Esquire means a lawyer in the US

@0celo7 I can always ask Reb for it :P

We exchange the occasional music piece

no, I found it

7:03 AM

Lawyers are slime balls in the US :P

David Z

@skullpatrol well now that's not very nice

Even if it is a widely held opinion

please remove it @DavidZ :-)

David Z

Eh, maybe

nvm

:-)

7:40 AM

anyway I apologize to all the lawyers in the US

user228700

8:25 AM

@JohnR: Look at my works ye mighty and despair:

user228700

;_; Agh, never mind, it is taking far too long to upload.

:: John puts the despair on hold while the picture is uploading ::

user228700

(I was, to kill the suspense, going to show you one of my Engineering Drawings)

Just as well really since I would have absolutely no idea what to make of an engineering drawing

@Sid perhaps, it's going to be apart of winter bash? Nuke the bots for prizes :-D

user228700

8:30 AM

Hahaha, right.

In fact I have little appreciation of any of the visual arts. To me art galleries are just places to be bored.

(apart from sculptures made from laptops of course)

@user685252 but, then that will be biased towards mods. Maybe, we get prizes for bringing spammers to their attention?

...good idea

user228700

user228700

Aha!

user228700

8:34 AM

(It is not very bright-I apologise)

What kind of pencil did you use?

Hi, third and @Loong (that's a football joke :)

user228700

@user685252 0.5mm HB and H lead pencils. Why dyou ask?

just wondering...

@Kaumudi.H looks fine to me ...

user228700

Fine?! Gosh, that took me about 25 minutes to draw!

8:42 AM

I did say I know nothing about engineering drawing ...

user228700

:-) I kid.

I think ...

Loong

@JohnRennie Yeah, looking at isometric projections of our plant piping gives me a headache.

user228700

Hahaha, funnily enough, learning to draw isometric projections is exactly what I am doing at the moment!

I suppose I've always regarded engineers as some variant of troglodyte, so I've never been that concerned with what they draw on their cave walls ...

8:50 AM

@JohnRennie Where did the wave equation $y(x,t)= y_m \sin (kx-\omega t)$ come from?

Historically it was derived as the equation of motion of a stretched string.

@JohnRennie Could you help me derive it?

I forget the details of the derivation, but basically you consider the string as made up from a sequence of point masses and work out the forces acting on each point mass.

there are loads of derivations a mere Google away

@JohnRennie I can't find the equation I wrote in any derivation.

There are two steps:
(1) derive the differential equation that describes the motion
(2) solve the equation

Your equation $y(x,t)= y_m \sin (kx-\omega t)$ is step (2) i.e. it's the solution to the differential equation.

8:58 AM

@JohnRennie Okay. For the time being please let me know the significance of $\omega t$ and $kx$.

The differential equation is: $$ v^2 \frac{\partial^2y}{dx^2} = \frac{\partial^2y}{dt^2} $$

That's just the one dimensional wave equation.

@JohnRennie Yes, I saw it's derivation in Walter Lewin's lecture. Do you know how to solve that?

If you substitute your equation into the differential equation you should see that it satisfies it.

I'm afraid it's an ansatz again :-)

:/, I hope I get used to ansatz soon.

Okay.

you will pal

eventually

9:02 AM

As an impressionable young physics student it was a disappointment to me to find that most differential equations are solved by guesswork. But with time you learn to love the ansatz :-)

Anonymous

There are standard rigourous methods to solve the wave equation. I don't know why they don't teach it in physics courses. :/

Because it's a whole lot of analysis

I don't even remember what's the most general solution to the wave equation

It's some set of the Sobolev space or something

Whole Lotta Analysis - one of Led Zeppelin's less well known tracks

Anonymous

@Slereah True. But I hate it when they say the only method to solve the wave equation is guesswork. They can at least refer sources which the students can look up to learn the standard methods. I had found a very nice one few months back. Gotta search for it. @BalarkaSen might know (?)

Well the wave equation is easy enough to solve in 2D

It's like $$\frac{\partial^2 f}{\partial u \partial v} = 0$$

Anonymous

9:09 AM

Yeah 2D one is there on the MIT pdf also I think

It's pretty obvious to solve that way

Anonymous

Here's one which uses separation of variables (though that is possible only for a certain type of functions). google.co.in/url?sa=t&source=web&rct=j&url=https://… . You could check it just for the motivation :P @Abcd

Well, the 2D case without considering the class of functions is pretty trivial :p

Dr.@JohnRennie In the mood for another question? (Mechanics-Forces this time) O:)

@Alan oh go on then :-)

You don't really need to call me Dr.

9:20 AM

:O

So, "Mr." was okay then? :D

I put the Dr on my driving licence just in case it helps me get off a speeding ticket one day, but apart from that I've never used it :-)

@Alan just John is fine

(Question's on pulleys)

@JohnRennie Alright :D

Something obvious must be evading me again... I'm told that the tension $T'$ is twice the tension $T$

The tension in the string acts upwards on the mass, but it acts downwards on the pulley. So the total downward force on the pulley is $2T$

@John Thanks, again :3

No problem, that was an easy one :-)

9:26 AM

At least you don't seem bothered by my constant (trivial) bugging... I'm relieved.

The reason students are given these sorts of problems is to build up a library of things you know just from a quick glance at a problem. That saves time when in later life you have to do these sorts of calculations for real.

Judging by my (limited) interaction with you... you're quite good at explaining stuff (lucky I ran into you) O:)

There are loads of us who can explain this sort of problem, but most are either revising for exams or working. My only special qualification is that since I'm retired I'm usually hanging around looking for something to do :-)

^ Read up on your profile... you've definitely been busy around here :D

Almost a quarter of a million reputation :O

@John Thanks for helping me out (yet again). I'll be off now!

Anonymous

9:46 AM

@JohnRennie "Silly people! I am a doc and I was overspeeding to save someone's life!" The cops would immediately let you go :P Good reason to do a PhD. :D

I'm a doctor and I need to solve a PDE!

I'm a doctor and I need to do a Dehn surgery on a 3-manifold

@Blue no, that would get you a blue light escort, then a speeding ticket plus a conviction for wasting police time :-)

@BalarkaSen Heyooo

hey hey hey

9:49 AM

That manifold got a Dehn twist in it, I must operate immediatly!

Mhm, it's got a rather bad complication along the tubular neighborhood of that link

It has a compact bowel

Sai Krishna Deep

10:14 AM

Hi, I'm looking for some good CMT with a good learning curve for motivated students who are about to enter this field for research. Well, I have pretty decent knowledge in QFT. Taken 2 graduate courses which covered the most essential stuff upto Renormalization, Spontaenous Symmetry breaking, Non-abelian Gauge theories, asymptotic freedom.. etc..

I like good theoretical books rather than hand wavy physical arguments and interpretations. But as far as I can tell, physical understanding is required in CMT more so than in QFT. Which is why I would request you to suggest some introductory book…

CMT?

Conformal milk theory?

Sai Krishna Deep

10:32 AM

Condensed matter

CooperCape

Is there any relation between the differential of a complex number and the differential of its conjugate?

maybe like $\frac{dz}{dx}=-\frac{dz*}{dx}$? Obvs I'm being fanciful here...

$dz/dx$ is a nonsensical expression

CooperCape

that's true...

umm

$\frac{dz(x)}{dx}$

what is $z(x)$

CooperCape

I'm always one for nonsensical expressions when I don't think.

Okay that question was ridiculous beyond belief (leaves through embarrassment)

10:37 AM

Well no but you need to know what a complex derivative is

Look up Wirtinger derivatives

@SaiKrishnaDeep Sorry condensed matter is too useful to society

I only do useless fields

Sai Krishna Deep

@Sle

@Slereah its okay, thank you for ur time.

Although

I still should try to do that linear gravity condensed matter idea I had

(I wonder what the metric generated by an infinite lattice is)

Sir Cumference

11:35 AM

Morning

11:56 AM

mroing

Akoben

Can someone take a look at my question and tell me if there is something wrong with it? physics.stackexchange.com/questions/372959/…

Got a downvote but not sure why....

infinite countable or infinite uncountable, though I guess rarely do people need to use uncountable ordinal labelled lattices

12:17 PM

It's a lattice

take a guess

uncountable**?**

No

A lattice is just $\mathbb{Z}^n$

Well, I don't see why you cannot call a cartesian product of ordinals a lattice, but perhaps I am missing some precise things a lattice need to obey besides periodicity.

But anyway, what I just said is not relevant to the discussion anyway, because we are clearly dealing with integer lattices.

So... you mention doing linear gravity on a lattice, are you trying to model the gravitational interactions of the atoms in the lattices of condensed matter systems, or some other quasiparticle states?

12:33 PM

Looks like the superuser spam have spilled to us

2

search: insurance

:O

9 hours ago, by user685252

What's going to happen to Winter Bash if they take over?

simple, they stole all our hats