« first day (3766 days earlier)      last day (1258 days later) » 

6:10 PM
Where can I find a proof of the statement:
 
hey hot cats
 
nullspace(A)=nullspace(Reduced Row echelon Form of A)
?
 
@satan29 you just have to show that the row operations preserve the nullspace.
This is actually a great exercise.
 
Oh! In that case, let me get my hands dirty
 
Let x be in null(A) and show that x is in null(A') where A' is obtained from A' by a row operation.
Let me know if you need a little push or hint.
 
6:21 PM
Now that i think.....its very intuitive but what I am thinking hasn't got any rigor to it
Consider the equation AX=0 as a system of linear equations'
E1=0
E2=0
..and so on
say there is a solution set x=(x1,x2...)
now, x satisfies E_{i}=0 (for i =1,,2....) so it also obviously satisfies k*E_{i}=0
for a constant k
 
I think that is enough rigour for that part.
 
If $E$ is invertible then $Ax=0$ iff $EAx=0$.
 
Though of course you have done only one of the three elementary row operations. So the two others remain.
One of them is stupidly simple.
 
and similarly, x satisfies E_{j}=0 so x will satisfy a*(E_{i}) + (b*E_{j})=0 for constants a and b
@anakhro i believe thats the interchanging two rows thing?
 
Indeed. :)
So you showed null(A) is a subset of null(A'). What about null(A') being a subset of null(A)?
 
6:29 PM
hmm
 
Why not use the ERO approach?
 
i cannot directly say that , if x satisfies aE1 + bE2=0 then it satisfies E1=0 and E2=0
 
Copper is here procrastinating another day! :)
 
Does anyone know the current method of (on the fly) inviting a user to a math chat? My rep is over 1000 and the user I wish to invite is at 21 rep. The old meta-math articles don't seem to work anymore.
 
@satan29 a hint: A' was obtained from A by a row operation...
 
6:33 PM
it has got to be just reversing what i did earlier, no? I cant seem to express it properly though
 
@TedShifrin I am an expert at procrastination. Unfortunately I don't bill those hours :-).
Good morning!
 
Good morning.
 
At least in Albany California :-)
 
its 12:04 AM in India lol
 
Used to be named Ocean View but there were so many Ocean View's in California the mayor at the time renamed it Albany. I prefer Ocean View albeit Bay View would be more truthful.
 
6:35 PM
Crazy half hours.
 
Nepal is even stranger :-)
 
It was Albany in the 50’s when I grew up there. Bay View, maybe.
 
Wow? It must have been a vastly different place then.
 
@satan29 Well in the easiest case, A' being A except with two rows swapped: how do I get A from A'?
 
Grew up in Kensington.
 
6:36 PM
Nice. Still nice.
 
@anakhro swap em again?
 
Yes :)
 
Yup. I go and drive by the house every time I am back visiting.
 
Berkeley, Kensington & El Cerrito are all about a block from my (the bank's) house.
 
But you already showed that swapping the rows doesn't change the null space!
 
6:37 PM
yes!
 
Wave when you pass Curtis & Thousand Oaks :-). Come in for a glass, hopefully Covid will have dissipated then.
 
I was in upper Kensington — top of Willamette.
 
the other two operations remain though
 
I pass through there a few times a week on my bike. Used to run but need a replacement hip apparently. Now there is a topic for procrastination.
 
Good thing you have such well-established expertise.
 
6:39 PM
My house was owned by a 3rd gen. member of the Jacuzzi family (of hot tub fame).
One of them dropped by last Christmas to see the old house :-).
 
So you're basically on Colusa.
Maybe TMI for the web, but I know precisely.
 
isnt that....too much info to be posting online?
 
Ironically my copper.hat name came from my siblings who used to make fun of my privacy concerns :-).
 
Yes. Remove it.
 
A quick google search starting from my MSE bio will yield the same result :-).
But I will remove it if it makes folks uncomfortable.
 
6:42 PM
I AM SO UNCOMFORTABLE.
it's cold outside. :(((((
 
Shaddup.
 
Its gone, sorry, didn't mean to induce discomfort.
 
No big deal. There are crazies around here.
I'll drop by if I ever get to travel again.
 
@copper.hat i was just joking. :(
 
:-) Free glass of wine, get your ears talked off :-)
 
6:45 PM
@anakhro halp... I am stuck :(
14 mins ago, by satan 29
i cannot directly say that , if x satisfies aE1 + bE2=0 then it satisfies E1=0 and E2=0
 
Deal, copper :)
 
@satan29 do you see how the row operations can be written as premultiplication by particularly simple matrices?
 
@satan29 how can you undo the row operations?
 
Yes, reversibility is the whole point of row operations.
 
how about this: since x satisfies aE1+bE2=0 for all (a,b), x MUST also satisfy the case where (a,b)= (1,0) and also the case where (a,b)=(0,1)
so we can recover E1=0 and E2=0
 
6:49 PM
Let's just make it kind of explicit what the row operation is. We have three elementary row operations: swapping rows, multiply rows by a non-zero scalar, and adding a multiple of one row to another.
You agree we are done the swapping rows case?
 
yes
multiplying with constant is also obvious
 
How is it obvious? Can you explicitly explain it?
 
just..multiply the eqn with the multiplicative inverse?
 
Great, and that again is a row operation, right?
 
Just a word of warning for when you’re done. In your earlier argument, $a$ and $b$ are fixed, not arbitrary.
 
6:51 PM
So our previous direction holds for null(A') being a subset of null(A).
 
yes.
@TedShifrin oh hmmmm indeed
 
So now suppose I add k times row Ei of A to row Ej. So the jth row of A' is aEi + Ej.
Given A', can you now "undo" it to A again?
 
subtract Ej , and multiplying by 1/a ?
 
That would mean the jth row is Ei.
(in A, it is Ej)
 
ok...
 
7:04 PM
What is the ith row of A'?
 
Ej?
and then we swap?
 
Okay, maybe this calls for some TeX
 
let $r_k$ be the $k$th row of $A$.
 
okay
 
Using copper.head's notation instead of Ei, Ej, etc.: $A = \begin{bmatrix}r_1\\\vdots\\r_n\end{bmatrix}$
 
7:09 PM
What are the formulae for the rows of $A'$?
 
@anakhro yes
 
As copper.hat asks, what are the rows of A'?
 
a1r1 +a2r2.... as the first row
b1r1+b2r2....2nd row
like this?
 
Sorry, what are you trying to do exactly?
 
we are trying to prove the part where you add multiples of rows together, right>?
 
7:16 PM
Yes, so suppose we obtain A' from A by adding $k$ times $r_i$ to $r_j$.
What are the resulting rows of A'?
(keep in mind this is a single row operation, not many).
 
r1....ri,(rj+kri)....rn
 
So, if $r_1,...,r_n$ are the rows of $A$ then the rows of $A'$ are something like $r_1,..., r_{j-1},r_j+ k r_i, r_{j+1},... , r_n$.
 
Yes, exactly, so we get $A' = \begin{bmatrix}r_1\\\vdots\\r_j + kr_i\\\vdots\\r_n\end{bmatrix}$.
So now we want to go backwards from A' to A by using row operations.
How do we "undo" it?
 
@satan29 your move :-)
 
subtract kri from the jth row?
 
7:21 PM
Why a question mark? Does that seem like it would be a problem?
 
no lol
 
Is it a row operation?
 
yes, adding (-k)ri
 
And you double checked that $r_i$ was another row of A'?
 
yes
 
7:23 PM
So everything checks out? Does the argument follow easily after that?
 
yes, indeed :)
this was so simple man......I just wasnt thinking in terms of single operations
 
@LukasHeger I just read your proof, thanks for writing that. It's nice.
 
When you have time, write down the form of the matrix corresponding to the above row operation.
 
Thank you so much @anakhro and @copper.hat for being so patient.
 
The expression as the product of row operations is useful in many contexts so is worthwhile remembering.
@anakhro did the heavy lifting
 
7:25 PM
@copper.hat which expression?
 
the reduction of $A$ to reduced row echelon form can be done by premultiplying by a sequence of elementary row (operation) matrices.
 
(an elementary row operation is an invertible linear map, so it also has the form of a matrix, which is not too hard to write down.)
 
cool
 
nice, but i can't say its immediately obvious to me how the colours are attached...
 
7:53 PM
I haven't really looked at it but it's HSV rather than RGB
Or actually I think they're both available
 
"When you procrastinate so much that you watch a video about what you are procrastinating"
me when I read expository texts about a topic that I should learn in detail or do exercises for
 
How do I get the bound $\vert\operatorname{SL}(2,\mathbb{Z}/N\mathbb{Z})\vert\le N^3$? A naive count only gives me the upper bound $2N^3-N^2$.
 
Maybe I should get bounds on how many inverses an element can have in Z/NZ, but I'm not seeing anything good
 
I can't help but read these in schwarzenegger-like german accents
 
8:04 PM
ok that does make them funnier
 
"commentate" is such a german error
 
doesn't he have an Austrian accent?
 
@Thorgott You can count cardinality of GL_2(F_q), SL_2(F_q) is the kernel of the determinant map so you just divide by p to get that
SL_2(Z/n) is a product of SL_2(F_q)'s right
 
@copper.hat different dialects are everywhere in german-speaking regions, the (terrible) german accents in english don't differ too much though
 
i like the variety. if only i could speak my native language...
 
8:07 PM
may I introduce to our proudest export product, and prime german meme?
 
SL_2(F_q) has cardinality 1/2(q^3 - q)
if I am not mistaken
So you get your bound
Much better than that
 
well, Z/n is a product of F_qs as Z-modules, so End(Z_n) is a product of End(F_q)'s as rings and they have the same invertible elements, but why does this respect determinants
 
yeah whatever screw it just say GL(Z/n) = prod GL(F_q) which has cardinality <= prod q^4
so at most n^4 and then uhh
yeah that doesnt quite work. interesting
 
GL(Z/n) having cardinality at most n^4 is obvious lol
but I mean, SL(Z/n) has index n
so you were right and that's it
 
So? lol
algebra brain
Also no man SL preserves product
its clear
 
8:11 PM
why
 
@user2103480 You Germans are so boring, this is what happens when an Italian politician embarasses himself speaking English at an international thing
 
My ears hurt
 
@AlessandroCodenotti omfg he's so f*cking italian
 
@Thorgott SL(R) -> SL(I) is surjective and has kernel SL(R/I) for nice rings like Euclidean domains or whatever by
row operations
elementary generation my dude
 
give a more abstract proof
 
8:14 PM
so SL(Z/n) \cong prod SL(F_q) and you get exact cardinality using my argument
 
@AlessandroCodenotti is there a strong expectation for Italian politicians to speak english?
 
Some guy made a video of the french equivalent youtu.be/AxkBqW4dK-E
 
@Thorgott Follows easily by cocompleting the appropriate category
 
thanks, that's much clearer
 
Also SL is q^3 - q, PSL is SL/2, of course
So N^3 prod_{q|N} (1 - 1/q) is the cardinality of SL_2(Z/N)
Number theorists come at me
 
8:15 PM
@Astyx lol
Some guy made a techno remix of the rector of my uni awarding degrees
 
You can be do what we want to do
 
@BalarkaSen I meant Euclidean rings not domains
I is a proper ideal
@BalarkaSen *q^2
 
user486313
Hi @MatsGranvik
 
8:35 PM
@BalarkaSen This is totally wrong man
PSL is of course SL/(1+(-1)^{q+1})
 
Lol
I lol'd irl
 
8:48 PM
Silent Hill 1-4 soundtracks are fantastic
brilliant games, wish i had the time and patience to set up and play SH2 on PC
 
Never been a fan of the games
The Kojima/Del Toro one looked super promising though
 
yeah
thatd be too high end for me to play anyway lol
i wasted my childhood man, should have played way more games
 
I waste my childhood, should have read more Lurie and less novels
 
lol
 
@BalarkaSen do you know Return of the Obra Dinn? A fellow PhD student suggested it to me and it looks great from the little I've seen (but I'm trying not to see much because I prefer playing games blind)
 
8:53 PM
once controllers had more than 3 buttons games became too complex for me
 
@AlessandroCodenotti nope
 
tfw you're grading HW and someone did the right problems in the wrong chapter
so close and yet so far
 
ohnooo
 
9:10 PM
@balarka i'd pity them more but
so many of these students just end up copying from the solutions manual
 
can someone help me understand this? $\frac{dxdy}{xy(x-1)(y-1)}$
 
that's...not a question
 
I mean this, $ds^2=\frac{dxdy}{xy(x-1)(y-1)}$
 
Hello everyone. I would like to know what you guys think of "Do every proof yourself first".
Any of you have studied as this during undergrad or grad school?
 
@geocalc33 once again, that's not a question.
 
9:15 PM
I mean this metric $ds^2=\frac{dxdy}{xy(x-1)(y-1)}.$ I'm trying to understand it better
when x=1 and y=1 then it's undefined
and when x=0 and y=0 it's undefined
 
that seems pretty dubious as a metric
pretty sure a valid first fundamental form has $ds^2=E dx^2+2F\,dxdy+G dy^2$ with $EG>F^2$. (maybe $\geq$, but not $F^2>EG$)
 
sorry I mean for a semi-riemannian space
 
ah. dunno in that case
 
@AttractorNotStrangeAtAll Different folks have different approaches. My preference is to 'understand' the result first and often the proofs will follow from this 'understanding'. However you can spend many unproductive hours trying to reproduce results that took other years to develop without having a road map at least.
 
9:33 PM
Hmmm, so maybe some drawing, intuitive reasoning, then the proof will probably easily follow? Unless it is dependent on some crazy trick obsly
 
well
if it took years for someone to come up with the argument, it may take years to come up with it again
 
But the difference is that you know it's true, right? But I get your point
 
the point in my book is that you never try to learn anything "from scratch". you always make use of the scaffolding that's been left behind
obviously, the more you can reconstruct without relying on said scaffolding, the better
but there's always a limit to that.
 
Right! So I should concentrate my study hours doing the proposed problems/homework
Sorry if all these questions sound a little "I don't know how to study", but I'm really learning now how to learn
 
right. you're trusting that whoever wrote the problems/HW knows the material well enough that they can judge the right path
that gets tricky when you're reading a textbook independently, of course
which you do end up having to do at some point
 
9:40 PM
Sometimes I understand the theorems, I 'get' the concepts and definitions, I'm ok with the examples, but when I arrive at the exercises I feel like there's something there I don't really understand what's going on. So that's why I thought about changing my study strategy
 
@Semi o/
 
Or maybe, as I'm new to math, this all is normal and it's actually good because it means I'm being challenged
 
@astyx \o
 
how are you?
 
alright. hopped on to write up a physics stack question (my first in years) and thought i'd stop by
 
9:43 PM
What are you doing nowadays?
 
"The Nielsen–Schreier theorem in turn implies a weaker version of the axiom of choice, for finite sets."
Is Wiki trolling? Choice for finite sets is provable in ZF already or am I tripping?
 
mostly some teaching type work. intro physics stuff
 
@Thorgott not for all families of finite sets, not even for countable families of finite sets (that should be equivalent to what logicians call the weak Koenig's lemma)
ZF doesn't even prove choice for families of unordered pairs iirc
 
oh, it's choice for collections of finite sets
I was thinking finite collections
 
right now i'm working on grading HW, and it's dull as anything
 
9:46 PM
yeah ok
 
@Thorgott Ah I see, that's provable in ZF, you're right
 
@AttractorNotStrangeAtAll I have found that having someone to discuss theorems, etc. with to be of great value.
 
@copper.hat with everything remote is harder, and I'm doing classes as a non-degree student, so no one is really my friend
I'll try to record some videos of myself explaining some concept or whatever, because it makes me explain things out loud and clear my thought of what's happning, Idk
With a drawing tablet, of course
 
yeah, independent learning in a math field you're not comfortable with is tough
 
@AttractorNotStrangeAtAll I understand. I have only found one person in my career (I'm 60) who shares a similar perspective & interest in mathematics and with whom I would socialise independently of mathematical motivations.
Part of doing mathematics efficiently involves a directness that makes some uncomfortable.
 
10:01 PM
does the sum of two metrics always result in a metric?
yes
this means the sum of 4 metrics is a metric
 
two positive-definite metrics, i'd say yes
not sure about beyond that, though. if A and -A are both permitted as metrics, then A+(-A)=0 is degenerate
which seems dubious
 
oh I overlooked that
 
How can $-A$ be a metric?
 
semi-Riemannian context
 
Of course, thanks.
 
10:05 PM
calling it a metric in that case does seem weird tho
 
well at least the ds^2 I gave isn't degenerate :)
 
Hi. Does anyone know where I could find a proof/explanation of this answer about how to Calculate 3D Vector out of two angles and vector length?
https://math.stackexchange.com/questions/1385137/calculate-3d-vector-out-of-two-angles-and-vector-length
 
@jarlemag The two angles define two planes and the intersection defines a line. Then you need a length and direction (plus or minus).
 
the description in that question is a bit strange, i think. it seems that $\alpha,\beta$ correspond to the angles of the vector upon projecting to the $XY$ and $YZ$ planes respectively
hmm, no, that doesn't work either
 
@Semiclassical Why not?
 
10:11 PM
@copper.hat take the (X,Y,Z) point to be (1,0,0). then the projection into the YZ plane is (0,0)
whereas the conversion given in the answer would require $\alpha=\pi/2$ and $\beta=0$. that makes sense for spherical coordinates but not for the question itself
(unless, of course, the OP's description of their angles was so unclear that it was actually correct)
 
@Semiclassical yup. need more caffeine :-)
 
hah
@jarlemag my reading of this is that, while that answer is fine for what it sets out to do, it doesn't really match the question as stated. but the lack of detail means i'm not really sure i understand what the OP asked for in the first place
 
user486313
hi @copper.hat
 
Hi @user131585!
 
user486313
does anyone read the work of Peter Lax?
 
10:16 PM
I got a bit confused by the question (and a similar one here: stackoverflow.com/questions/30011741/…) too, but the basic question is still "how to get (unit) vector from two angles."...
 
Yeah, something seems odd.
 
I think this would be a statement applicable to my case: "α is an angle in the XY plane (angle from the X axis), β is an angle between the XY plane and the Z axis. Find the unit vector with direction defined by the angles.
 
@jarlemag right. That description matches the first answer.
it doesn't match the second question you linked, though, which makes me a bit dubious of the latter
 
Hello chat!
 
user486313
Hello
 
10:24 PM
yeah, the second question doesn't seem to be answered correctly
@jarlemag That said, it's not too hard to validate the first question's answer in your case
First, if you project (X,Y,Z) into the XY plane, you get the point (X,Y)
 
I have a silly question: If we know that $|A|\leq|B|$ for all sets $A,B$, can we find an example of $|A|>|B|$?
 
user486313
I'm going all-out tonight for dinner -- steak burrito, belgian fries, and ice cream ...
 
user486313
then, write a research statement ...
 
@manooooh I don't understand the question. What does $|A|\leq |B|$ for all sets $A,B$ mean?
Surely this means $|A$ is a constant for the collection of sets in question?
 
@copper.hat oh sorry. $|A|$ stands for cardinality of the set $A$
 
10:27 PM
I got that, that was not what I was concerned about.
 
@manooooh that's no better. what's your domain for A,B?
i mean, |A|>=|B| is not true for arbitrary sets A,B
so asking what the consequences of it being true is...odd
 
I rephrase: Given two sets $A,B$ that satisfy $|A|\leq|B|, can we also have $|A|>|B|$?
 
Sure. Take A to be any non-empty set and B the empty set.
n>=0 and n>0
 
He's not asking for that
He wants 0>n
 
10:30 PM
@Semiclassical sorry, I must specify that both $A,B\neq\emptyset$
 
What does < mean tho?
 
then...no? i mean, $x\leq y\leftrightarrow x\not >y$
 
@Semiclassical I see, one is the negation of the other. I guessed it was a silly question. Thanks!
 
It depends how you define the relations tho
 
this is for |A|<|B| as numbers, to be clear. it's not so clear if you mean A<B
 
10:32 PM
Yes of course, $|A|$ stands for "Amount of elements in the set $A$". And $<$ is the relation "less than"
 
So is |N| < |R| ?
If you're only dealing with finite sets, then this result is just the order on natural integers
If not, that definition isn't sufficient
 
Let me explain with one example
I am given a formal grammar $G=(V_n,V_t,P,S)$ where $V_n$ is the set of non-terminals (variables), $V_t$ is the set of terminals (alphabet), both are finite, $P$ are the productions or grammar rules and $S$ is the start symbol
I am asked to prove or disprove $|L(G)|=|V_t^*|$, which is obviously false
I have shown that $|L(G)|\leq|V_t^*|$, where $L(G)$ and $V_t^*$ can be finite or infinite
Is this sufficient to show that $|L(G)|\not>|V_t^*|$ hence $|L(G)|\neq|V_t^*|$?
Definitions of formal Language and Grammar can be found at its.caltech.edu/~matilde/FormalLanguagesGNSlides.pdf
 
replace those symbols with words and then ask yourself that question again
 
@Semimclassical: I tried to solve it this way before reading the questions/answers mentioned, but it doesn't seem to give right results. If anyone can see something obviously wrong, I'd be happy to learn where I misstepped:

α is an angle in the XY plane (angle from the X axis), β is an angle between the XY plane and the Z axis. Find the corresponding unit vector v.

We know that:
x^2 + y^2 + z^2 = 1

Let f be the length of the projection of v into the XY plane. Then:

sqrt(f) = sqrt(x^2 + y^2)
Uh, I hope I remembered that trignometric identity right...
 
@Thorgott yeah but some symbols need to be put there, isn't it? I mean, the generated language can be equal to the set of all words of any length (this is what $V_t^*$ means), but not always. Of course it can occur that $|L(G)|<|V_t^*|$, but my question is: How can we explain that it is impossible -no matter what grammar is- that $|L(G)|>|V_t^*|$?
So my question, again, is: Is it sufficient to prove that $|L(G)|\leq|V_t^*|$ to conclude $|L(G)|\not>|V_t^*|$?
 
10:47 PM
@jarlemag I agree with what you have for x,y but not for z
 
what I'm saying is that you should remind yourself of what those inequality symbols mean in plain language and then it will be very clear which parts of reasoning are sound and which aren't
 
@Semiclassical Interesting! Can you give any more clues? :-D
 
you're asking "if $L(G)$ is at most as big as $V_t^{\ast}$, can it be greater than $V_t^{\ast}$?"
 
well, you have x^2+y^2=f^2 which is fine
but in that case, 1=x^2+y^2+z^2=f^2+z^2
so should have z=+- sqrt(1-f^2)
 
@Thorgott of course not
 
10:50 PM
indeed
@manooooh now try to see if you can figure out whether the "hence" here is reasonable or not
 
@jarlemag so what remains is to figure out what f and z need to be to give you the desired angle beta, subject to f^2+z^2=1
 
Thanks for the help! I'll take a look.
 
but that's not too hard either: should be be able to use trig on the right triangle formed by f,z and 1
 
@Thorgott it is not reasonable. The correct should be: Since $|L(G)|\leq|V_t^*|$ but $|L(G)|\not>|V_t^*|$, we conclude $|L(G)|\neq|V_t^*|$
Now?
 
Hello. I'm looking to ask a question concerning examples of finite concepts that require infinity. I'm not sure how it would be received. It might be too broad, say, or not phrased in a helpful way. Do you have any advice? What are your thoughts on some examples in the spirit of the question?
 
10:53 PM
nope, replace inequality signs with the words "greater" and "smaller" and think
 
that seems pretty broad yeah
 
I feel like I've seen such a question before
 
Do you have an example in mind to start with?
 
Some finitely presented groups are infinite.
 
@Shaun I don't know much of this, but an example that comes to mind is the methods used in the study of finite simple group of Lie type. These can be described as $G(\Bbb F_q)$ where $G$ is an algebraic group. It turns out that it's often useful to work with $G(\overline{\Bbb F_q})$ when studying those, e.g. their representations
 
10:57 PM
My motivation is to have something to point to in discussion with people skeptical of the rôle of infinity in mathematics.
 
@Shaun i feel like the idea here would be that certain infinite structures can be described using a finite set of rules
 
f(x)=x can be described using f(x)=x but it's actually an infinite object
 
but that is really general. for instance, a recursion relation generates an infinite sequence using a basic set of ingredients
 
That's the sort of thing I have in mind, @LukasHeger; thank you.
 

« first day (3766 days earlier)      last day (1258 days later) »