Explore
For Enterprise
Join for Free
Log In

Kunst und Geisteswissenschaften
Wirtschaft
Informatik
Datenverarbeitung
Informationstechnologie
Gesundheit
Mathematik und Logik
Persönliche Entwicklung
Physik und Ingenieurwesen
Sozialwissenschaften
Sprachen lernen
Abschlüsse
Zertifikate

Alle Bereiche erkunden

Blättern
Suche
Für Unternehmen
Anmelden
Kostenlose Teilnahme

Composition or combination of matrix transformations

Loading...

Mathematics for Machine Learning: Linear Algebra

Imperial College London

4.7 (3,208 ratings) | 78K Students Enrolled

Kurs 1 von 3 in Mathematics for Machine Learning Spezialisierung

In this course on Linear Algebra we look at what linear algebra is and how it relates to vectors and matrices. Then we look through what vectors and matrices are and how to work with them, including the knotty problem of eigenvalues and eigenvectors, and how to use these to solve problems. Finally we look at how to use these to do fun things with datasets - like how to rotate images of faces and how to extract eigenvectors to look at how the Pagerank algorithm works. Since we're aiming at data-driven applications, we'll be implementing some of these ideas in code, not just on pencil and paper. Towards the end of the course, you'll write code blocks and encounter Jupyter notebooks in Python, but don't worry, these will be quite short, focussed on the concepts, and will guide you through if you’ve not coded before. At the end of this course you will have an intuitive understanding of vectors and matrices that will help you bridge the gap into linear algebra problems, and how to apply these concepts to machine learning.

Lehrplan anzeigen

Skills You'll Learn

Eigenvalues And Eigenvectors, Basis (Linear Algebra), Transformation Matrix, Linear Algebra

Bewertungen

4.7 (3,208 ratings)

5 stars
2,349 ratings
4 stars
689 ratings
3 stars
119 ratings
2 stars
33 ratings
1 star
18 ratings

PL

Aug 26, 2018

Great way to learn about applied Linear Algebra. Should be fairly easy if you have any background with linear algebra, but looks at concepts through the scope of geometric application, which is fresh.

VC

Jul 11, 2018

The content of the course is very relevant, and the instructors are really fun and helpful.My only suggestion is to upload revisions for each assessment, so we can understand what we are doing wrong.

Aus der Unterrichtseinheit

Matrices in Linear Algebra: Objects that operate on Vectors

Now that we've looked at vectors, we can turn to matrices. First we look at how to use matrices as tools to solve linear algebra problems, and as objects that transform vectors. Then we look at how to solve systems of linear equations using matrices, which will then take us on to look at inverse matrices and determinants, and to think about what the determinant really is, intuitively speaking. Finally, we'll look at cases of special matrices that mean that the determinant is zero or where the matrix isn't invertible - cases where algorithms that need to invert a matrix will fail.

How matrices transform space5:41

Types of matrix transformation8:38

Composition or combination of matrix transformations8:59

Unterrichtet von

David Dye
Professor of Metallurgy
Samuel J. Cooper
Lecturer
A. Freddie Page
Strategic Teaching Fellow

Skript

Now, what's the point of telling about

all these different geometric transformations?

Well, if you want to do any shape alteration,

save all the pixels in

an image of a face or something like that,

then you can always make that shape change out

of some combination of rotations,

shear structures and inverses.

That is if I first apply

one translation A1 to a vector r,

then that makes some first change.

Then if I apply another translation A2

to the results of that,

then I can perform a composition of the two translations.

What I've done is I've performed first A1 and

then A2 to r. Now,

maybe this isn't so obvious,

so let's slow down into a concrete example.

Let's start out with our basis vectors.

So we've got e1 as being one,

zero, and e2 as being zero, one.

Now, let's take our first transformation

A1 as being a 90 degree anticlockwise rotation.

So what happens if we

rotate this by 90 degrees anticlockwise?

So e1 comes down to some transformed e1,

let's call it e1 prime of zero, minus one.

So put zero, minus one in there

and our e2 rotates down here,

to be some transform version of e2,

which will then be e2 primed,

which will be one, zero.

So then we've got an overall A1

which describes a 90 degree rotation.

Now, let's take another matrix

which also transforms our original basis vectors,

original e1 and e2.

What we'll say is that is a let's say a mirror.

So that moves a vertical mirror moves e1 to e1 prime,

is going to be minus one, zero.

It's going to leave e2,

where it was if I just reflect vertically.

So my transformation A2 now is going to be minus one,

zero, and it leaves the other one the same zero, one.

Now, let's ask ourselves what happens if I do A2 to A1.

So now I'm going to reflect over the result of doing A1.

So e1 prime is actually when I reflect vertically,

going to stay in the same place.

That's going to give me e1 let's say

double-prime for doing it twice.

E2 prime it's going to reflect over here,

and it's going to become e2

double prime is going to be minus one, zero.

So I'm going to have an overall result of doing A2 to A1,

which is going to be, well

I get e1 prime I first write down zero,

minus one and e2

double prime I'm going to write down minus one, zero.

Now, I can work out what that is,

actually in a matrix way without having to

draw it all out by saying,

A2 to A1 is doing A2 which is minus one,

zero, zero, one, to A1 basis vectors.

A1 was zero, minus one, one,

zero which is just the two

transform basis vectors the single primes.

When I do that, I just have to do this matrix to

that basis vector and then

this matrix to that basis vector.

What that looks like,

if I do that is I do

this matrix to that basis vector so I'll

do that row times that column.

That gives me minus one times zero

plus zero times minus one that gives me zero.

That one to that one.

So the second row first column zero

times zero plus one times minus one gives me minus one.

Then I do that A2 translation to the second column.

Now, to the second basis vector with A1.

So minus one times one,

plus zero times zero,

of minus one times one second row

zero times one plus one times zero gives me zero.

So I do the row times the column

for all the possible row and column combinations.

So that is in fact the same thing.

So we can have discovered really

how to do matrix composition,

or matrix multiplication doing

one matrix to another transformation matrix.

Notice that geometrically, we can show that A2 then A1,

isn't the same as doing the translations in

the other order first A1 and A2.

So just watch that for a minute.

So this one we've done A2,

and then what if we then do A1.

A1 is a 90 degree rotation so if we rotate e1 prime,

then we find that e1 double prime,

would be equal to our original e2 that's A1.

Our e2 primed, which was just staying where it was,

when we rotate that down,

then that'll come down to here.

That'll come down to one, zero,

so our e2 double prime will be one, zero.

Let's look how matrix wise.

So if we do A1 to A2,

A1 was zero, one, minus one, zero.

If we do that to A2,

which was minus one, zero, zero,

one that we're going to do that matrix multiplication.

What that's going to give us

is zero times minus one plus one times zero,

zero, then this one times this one,

minus one times minus one plus zero times zero is one.

This row second column zero, zero, one,

one and second row second column minus one,

naught, naught, one gives me naught.

So that is the first column is zero,

one, and the second column is one, zero.

So that isn't the

same as doing the operations in the other sequence.

You see these minus signs are flipped.

In fact what happened is,

this composition rotating and then flipping over,

is the equivalent of reflecting the whole lot,

the original basis vectors in this mirror,

and flipping and then rotating

round it effect is I've just flipped,

my original axis e1 and e2

which is putting a mirror in here.

So doing the two operations

in opposites in different sequences,

doesn't give you the same operations.

What we've shown is

that matrix multiplication isn't commutative.

A2, A1 isn't the same as A1, A2.

So we have to be very careful with matrix multiplication.

We can do them in any order meaning they're associative.

That is if we do A3 to A2, to A1,

we can do A2, A1 and then do A3,

or we can do A3, A2 to A1.

So we could do that, and then that.

Those are the same, they're associative.

But we can't interchange the order.

We can't swap them around.

It's not commutative.

So this is interesting.

There seems to be some deep connection

between simultaneous equations,

these things called matrices,

and the vectors we're talking about in the last module.

It turns out the key to

solving simultaneous equation problems,

is appreciating how vectors are transformed by matrices.

Which is the heart of linear algebra.

Durchsuchen Sie unseren Katalog

Melden Sie sich kostenlos an und erhalten Sie individuelle Empfehlungen, Aktualisierungen und Angebote.

Coursera arbeitet mit erstklassigen Universitäten und Organisationen zusammen, um Online-Kurse anzubieten und dadurch universellen Zugriff auf die weltweit beste Ausbildung zu ermöglichen.

© 2019 Coursera Inc. Alle Rechte vorbehalten.

Aus dem App Store herunterladen

Erhältlich bei Google Play

Coursera
Info
Leitung
Jobs
Katalog
Zertifikate
MasterTrack™ Certificates
Abschlüsse
Für Unternehmen
Für Regierungen

Community
Learners
Partners
Developers
Beta Testers
Translators

Verbinden
Blog
Facebook
LinkedIn
Twitter
Instagram
Youtube
Tech-Blog

mehr
Bedingungen
Datenschutz
Hilfe
Zugänglichkeit
Presse
Kontakt
Verzeichnis
Partnerunternehmen