[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]
2023-11: Warosu is now out of extended maintenance.

/sci/ - Science & Math


View post   

File: 117 KB, 694x854, 01-Numbers-1-Intro.jpg [View same] [iqdb] [saucenao] [google]
1115475 No.1115475 [Reply] [Original]

I wrote a tutorial on how to use diagrams to do computations with tensors.

I posted the beginning a few days ago - now I have more. I probably won't be posting any more updates here since it would be too long.

>> No.1115482
File: 188 KB, 846x1098, 02-Numbers-2-Grouping.jpg [View same] [iqdb] [saucenao] [google]
1115482

Tensors don't show up until page 15
Calculating components of a tensor is at the end

>> No.1115485
File: 119 KB, 715x845, 03-Vectors-1.jpg [View same] [iqdb] [saucenao] [google]
1115485

>> No.1115494
File: 154 KB, 730x1033, 04-Vectors-2-Rules.jpg [View same] [iqdb] [saucenao] [google]
1115494

>> No.1115501
File: 163 KB, 773x996, 05-Functions-1.jpg [View same] [iqdb] [saucenao] [google]
1115501

The flood blocker is lame...

>> No.1115505

>>1115475
WTF? do they use this shit anywhere? whats wrong with the standard tensor notation? with the subscripts and superscripts, it seeems pretty easy to me, at leasy easier then this!

>> No.1115507
File: 69 KB, 756x445, 06-Functions-2.jpg [View same] [iqdb] [saucenao] [google]
1115507

>> No.1115515
File: 167 KB, 744x993, 07-Covectors-1.jpg [View same] [iqdb] [saucenao] [google]
1115515

>>1115505
You of little faith.
You haven't even seen the notation yet! Wait until page 15

It's cake compared to the usual notation.

>> No.1115520
File: 63 KB, 691x557, 08-Covectors-2.jpg [View same] [iqdb] [saucenao] [google]
1115520

>> No.1115524
File: 79 KB, 748x708, 09-Vector Example-1.jpg [View same] [iqdb] [saucenao] [google]
1115524

>> No.1115532
File: 71 KB, 692x601, 10-Covector Example-1.jpg [View same] [iqdb] [saucenao] [google]
1115532

>> No.1115540
File: 177 KB, 720x1023, 11-Not Covectors-1.jpg [View same] [iqdb] [saucenao] [google]
1115540

>> No.1115546
File: 42 KB, 634x405, 12-Not Covectors-2.jpg [View same] [iqdb] [saucenao] [google]
1115546

>> No.1115555
File: 165 KB, 710x923, 14-Connect with Care.jpg [View same] [iqdb] [saucenao] [google]
1115555

>> No.1115551
File: 170 KB, 689x986, 13-Tensor Product.jpg [View same] [iqdb] [saucenao] [google]
1115551

I lied, tensors show up on page 13

>> No.1115558
File: 125 KB, 631x762, 15-Please Label.jpg [View same] [iqdb] [saucenao] [google]
1115558

>> No.1115565
File: 160 KB, 751x896, 16-Adding Tensors.jpg [View same] [iqdb] [saucenao] [google]
1115565

And here is one place where the notation really shines over the standard notation

>> No.1115568
File: 89 KB, 590x618, 17-Adding Tensors-2.jpg [View same] [iqdb] [saucenao] [google]
1115568

>> No.1115575
File: 130 KB, 651x912, 18-Contraction.jpg [View same] [iqdb] [saucenao] [google]
1115575

I don't think anyone will care if I go away for half an hour - this is a good place to take a break

>> No.1115577

For tensors, this notation doesn't seem to me to be much better than indices, but for spinors and for Lie algebra representations, it's wonderful. The Fierz identity is so much easier to use graphically.

>> No.1115584

MAKE MORE TUTORIALS
MAKE PDFS
MAKE

>> No.1115595

>>1115575
>>1115475

Yeah, this notation is bullshit. Im sticking to the standard tensor notation, its much better, neater, and understandable.

>> No.1115614

Thank you, OP. I've seen this notation sometimes, and never quite gotten it, so nice to read a good explanation. However, abstract indices works kinda fine too imo.

>> No.1115615

>>1115595
AGREED, WAY TO DO SOMETHING POINTLESS OP!

>> No.1115629

>>1115595
>understandable
Took me long enough to like index notation.

>> No.1115681
File: 118 KB, 730x972, 19-Zero Covector.jpg [View same] [iqdb] [saucenao] [google]
1115681

The rest of this is on how to compute components of tensors.

>> No.1115684
File: 99 KB, 675x836, 20-Zero Vector.jpg [View same] [iqdb] [saucenao] [google]
1115684

>> No.1115698
File: 175 KB, 1435x1747, 21-Generators.jpg [View same] [iqdb] [saucenao] [google]
1115698

>> No.1115703

>>1115475
rapidshit plox!

>> No.1115704
File: 194 KB, 1457x1880, 22-Linear Dependence.jpg [View same] [iqdb] [saucenao] [google]
1115704

>> No.1115712
File: 147 KB, 1484x1477, 23-Basis.jpg [View same] [iqdb] [saucenao] [google]
1115712

>>1115703
Eh, if you want it up on Rapidshare, you can do it.

>> No.1115715
File: 149 KB, 1484x1433, 24-Our Standard Basis-Components.jpg [View same] [iqdb] [saucenao] [google]
1115715

>> No.1115719
File: 177 KB, 1528x1753, 25-Characterizing Covectors.jpg [View same] [iqdb] [saucenao] [google]
1115719

>> No.1115723
File: 133 KB, 1490x1455, 26-The Dual Basis.jpg [View same] [iqdb] [saucenao] [google]
1115723

>> No.1115726
File: 206 KB, 1418x2007, 27-For Fun-The Cast.jpg [View same] [iqdb] [saucenao] [google]
1115726

>> No.1115736
File: 198 KB, 1424x2023, 28-Taming the Identity-1.jpg [View same] [iqdb] [saucenao] [google]
1115736

Dissecting a tensor with a Mathemagician Tool

>> No.1115745
File: 237 KB, 1462x2128, 29-Taming the Identity-2.jpg [View same] [iqdb] [saucenao] [google]
1115745

>> No.1115749
File: 216 KB, 1462x2078, 30-Taming the Identity-3.jpg [View same] [iqdb] [saucenao] [google]
1115749

>> No.1115754
File: 113 KB, 1539x1213, 31-Disecting a Tensor-1.jpg [View same] [iqdb] [saucenao] [google]
1115754

>> No.1115763

Cute tensors. Wouldn't have thought to encounter these.
I'll probably mail you in a few days once your tutorial is finished, I think it's an interesting alternative way of dealing with tensors. (... but I'm still wondering whether it has some practical use. Any info on that?)

>> No.1115774

OP, they have this thing called Powerpoint now.

You should look into it.

>> No.1115777
File: 198 KB, 1501x2095, 32-Disecting a Tensor-2-The Matrix.jpg [View same] [iqdb] [saucenao] [google]
1115777

And this is the last one.

The topics not included are derivatives and Lie Algebras. I don't have anything written on those yet, and it will probably be a while until I do. I probably won't be posting them here unless people really want them (but it seems like there are only a handful of people who would care)

>> No.1115802

>>1115763

If you're doing stuff with a whole bunch of tensors, lot's of contractions, etc, then the diagrams are much easier to look at and work with.
Of course there's nothing that you can't do with the original notation. I just like looking at diagrams much more than a long line of algebra.

>>1115774
I'd have to draw the diagrams by hand anyway. Powerpoint would just make this harder!

>> No.1115823

>>1115763
I've also posted this at PhysicsForums I'll be putting any new pages up there (new pages probably won't be for a while)

>> No.1115839

archive

>> No.1115960

>>1115763
I don't think this method will turn out to have any practical use. I think it'd be a great way to teach the concepts behind multilinear algebra. I could be wrong, though; a friend of mine started working on a diff.geometry notation that I never thought would work, but she's using it fine. In practice, I find alternate notations are more of a concept/sanity check than a way to do calculations or proofs.

>> No.1116395

>>1115960
This guy Predrag Cvitanović used the notation to find a new characterization of all of the Lie groups. Apparently they all fall out in some nice hierarchy in terms of invariants of the diagrams, even the exceptional groups.

There are definitely applications, but aside from nicer notation, they're pretty abstract stuff.

Of course Feynman diagrams are exactly tensors written in this notation and Physicists find them very useful.

>> No.1116530

>>1116395
idk about Cvitanovic's stuff, but I do know Feynman diagrams. At least in that case, I stand by what I said: Feynman diagrams are an awesome way to check concepts and sanity, but they don't necessarily make the calculations any easier. Certainly not until you've done the calculations enough to know what the Feynman diagrams are telling you in the first place...

But that's the caveat with new notation, isn't it? It's new. I'm skeptical that it'll work in practice, but that's because I haven't tried. And it'll take practice to get used to.

>> No.1117437

Awesome