Eigenvalues in a Nutshell

32
Eigenvalues in a nutshell Eigenvalues in a nutshell Mariquita Flores Garrido UDLS, March 16 th 2007

description

Although eigenvalues are one of the most important concepts in linear algebra, some of us eigen-struggle with them without understanding their usefulness and beauty. In this talk I'll briefly review the definition of eigenvalues emphasizing the associated geometric idea and I'll show how can they be used in some applications.From the Un-Distinguished Lecture Series (http://ws.cs.ubc.ca/~udls/). The talk was given Mar. 16, 2007

Transcript of Eigenvalues in a Nutshell

Page 1: Eigenvalues in a Nutshell

Eigenvalues in a nutshellEigenvalues in a nutshell

Mariquita Flores Garrido

UDLS, March 16th 2007

Page 2: Eigenvalues in a Nutshell

• Scalar multiple of a vector

• Addition of vectors

Just in case…

x

λx

xx x

λx

λx λx10 ≤≤ λ λ≤1 01 ≤≤− λ 1−≤λ

v1

v2

v1 + v2

Page 3: Eigenvalues in a Nutshell

Linear Transformations

mnnm RRfRA a:⇒∈ ×

V. gr.

• Rectangular matrices

⎟⎟⎟

⎜⎜⎜

635241

⎟⎟⎠

⎞⎜⎜⎝

⎛11

⎟⎟⎟

⎜⎜⎜

975

=

Ax = b Transformation of x by A.

n x 1

=A x Ax

m x n m x 1

Page 4: Eigenvalues in a Nutshell

*Stretch/Compression *Rotation

Linear Transformations

• Square Matrices

*Reflection

⎟⎟⎠

⎞⎜⎜⎝

⎛2002

⎟⎟⎠

⎞⎜⎜⎝

⎛0110

⎟⎟⎠

⎞⎜⎜⎝

⎛− ϕϕ

ϕϕcossinsincos

nnnn RRfRA a:⇒∈ × (*endomorphism)

Page 5: Eigenvalues in a Nutshell

*Shear in y-direction *Shear in x-direction

Bonnus: Shear

⎟⎟⎠

⎞⎜⎜⎝

⎛10

1 k⎟⎟⎠

⎞⎜⎜⎝

⎛101

k

x x

V.gr. Shear in x-direction

y y⎟⎟⎠

⎞⎜⎜⎝

⎛yx

⎟⎟⎠

⎞⎜⎜⎝

⎛ +ykyx

Page 6: Eigenvalues in a Nutshell

Basis for a Subspace

A basis in Rn is a set of n linearly independent vectors.

e3

e2

e1

2e3

⎟⎟⎟

⎜⎜⎜

211

⎟⎟⎟

⎜⎜⎜

211

⎟⎟⎟

⎜⎜⎜

001

⎟⎟⎟

⎜⎜⎜

010

⎟⎟⎟

⎜⎜⎜

100

= 1 + 1 + 2

Page 7: Eigenvalues in a Nutshell

Basis for a Subspace

Any set of n linearly independent vectors can be a basis

e1

e2 V1

V2

⎟⎟⎠

⎞⎜⎜⎝

2

1

aa

⎟⎟⎠

⎞⎜⎜⎝

⎛−=⎟⎟

⎞⎜⎜⎝

⎛12

2

1

aa

Using canonical basis:

??2

1 =⎟⎟⎠

⎞⎜⎜⎝

⎛aa

V1

V2

Using V1, V2 … ?

Page 8: Eigenvalues in a Nutshell

EIGENVALUES

•"Eigen" - "own", "peculiar to", "characteristic" or "individual“; "propervalue“.

• An invariant subspace under an endomorphism.

• If A is n x n matrix, x ≠ 0 is called an eigenvector of A if

Ax = λx

and λ is called an eigenvalue of A.

Page 9: Eigenvalues in a Nutshell

*Stretch/Compression *Rotation

Quiz 1

• Square Matrices (endomorphism)

*Reflection

⎟⎟⎠

⎞⎜⎜⎝

⎛2002

⎟⎟⎠

⎞⎜⎜⎝

⎛0110

⎟⎟⎠

⎞⎜⎜⎝

⎛− ϕϕ

ϕϕcossinsincos

Page 10: Eigenvalues in a Nutshell

• Characteristic polynomial: A degree n polynomial in λ:

det(λI - A) = 0Scalars satisfying the eqn, are the eigenvalues of A.

V.gr.

• Spectrum (of A) : { λ1, λ2 , …, λn}

• Algebraic multiplicity (of λi): number of roots equal to λi.

• Eigenspace (of λi): Eigenvectors never come alone!

• Geometric multiplicity (of λi): number of lin. independent eigenvectors associated with λi.

Eigen – slang

02543

214321 2 =−−=

−−

⎯→⎯⎟⎟⎠

⎞⎜⎜⎝

⎛λλ

λλ

)()( kxkxAxkAxk

xAx

λλ

λ

=⋅=⋅

=

Page 11: Eigenvalues in a Nutshell

Eigen – slang

• Eigen – something: Something that doesn’t change under some transformation.

xx

edxed

=][

Page 12: Eigenvalues in a Nutshell

FAQ (yeah, sure)

• How old are the eigenvalues?They arose before matrix theory, in the context of differential equations.

Bernoulli, Euler, 18th Century.

Hilbert, 20th century.

• Do all matrices have eigenvalues?Yes. Every n x n matrix has n eigenvalues.

Page 13: Eigenvalues in a Nutshell

• Why are the eigenvalues important?

- Physical meaning (v.gr. string, molecular orbitals ).

- There are other concepts relying on eigenvalues (v.gr. singular values, condition number).

- They tell almost everything about a matrix.

Page 14: Eigenvalues in a Nutshell

1. A singular ↔ λ = 0.

2. A and AT have the same λ’s.

3. A symmetric Real λ’s..

4. A skew-symmetric Imaginary λ’s..

5. A symmetric positive definite λ’s > 0

6. A full rank Eigenvectors form a basis for Rn.

7. A symmetric Eigenvectors can be chosen orthonormal.

8. A real Eigenvalues and eigenvectors come in conjugate pairs.

9. A symmetric Number of positive eigenvalues equals the number of positive pivots. A diagonal λi = aii

Properties of a matrix reflected in its eigenvalues:

Page 15: Eigenvalues in a Nutshell

10. A and M-1AM have the same λ’s.

11. A orthogonal all |λ | = 1

12. A projector λ = 1,0

13. A Markov λmax = 1

14. A reflection λ = -1,1,…,1

15. A rank one λ = vTu

16. A-1 1/λ(A)

17. A + cI λ(A) + c

18. A diagonal λi = aii

19. Eigenvectors of AAT Basis for Col(A)

20. Eigenvectors of ATA Basis for Row(A)

Properties of a matrix reflected in its eigenvalues:

M

Page 16: Eigenvalues in a Nutshell

What’s the worst thing about eigenvalues?

Find them is painful; they are roots of the characteristic polynomial.

* How long does it take to calculate the determinant of a 25 x 25 matrix?

* How do we find roots of polynomials?

Page 17: Eigenvalues in a Nutshell

WARNING:

The following examples have been simplified to be presented in a short

talk about eigenvalues. Attendee discretion is advised.

Page 18: Eigenvalues in a Nutshell

Example 1: Face Identification

Eigenfaces: face identification technique.

(There are also eigeneyes, eigennoses, eigenmouths, eigenears,eigenvoices,…)

Page 19: Eigenvalues in a Nutshell

EIGENFACES

Given a set of images, and a “target face”, identify the

“owner” of the face.

128 images

(train set)

256 x 256

(test)

Page 20: Eigenvalues in a Nutshell

1. Preprocessing stage: linear transformations, morphing, warping,…

2. Representing faces: vectors (Γj) in a very high dimensional space.

V.gr.

Training set: 65536 x 128 matrix

3. Centering data: take the “average” image and define every Φj

∑=

Γ=Ψn

jjn 1

1

jj Γ−Ψ=Φ

],...,,[ 21 nA ΦΦΦ=

Page 21: Eigenvalues in a Nutshell

4. Eigenvectors of AAT are a basis for Col(A) (what’s the size of this matrix?), so instead of working with A, I can express every image in another basis.

* 5. PCA: reducing the dimension of the space. To solve the problem, the work is done in a smaller subspace, SL, using projections of each image onto SL.

6. It’s possible to get eigenvectors of AAT using eigenvectors of ATA.65436 x 65436 128 x 128

Page 22: Eigenvalues in a Nutshell

Example 2: Sparse Matrix Computations

Page 23: Eigenvalues in a Nutshell

ITERATIVE METHODS

 x = b

• Gauss-Jordan

• If  is 105 ×105 , Gauss Jordan would take approx. 290 years.

• Iterative methods: find some “good” matrix A and apply it to some initial vector until you get convergence.

• Choosing different A determines different methods (v.gr. Jacobi, Gauss-Seidel, Krylov subspace methods, …).

Page 24: Eigenvalues in a Nutshell

Example 2: ITERATIVE METHODS

0n

n

02

012

01

xA x

xA )A(Ax Ax x

Ax x

=

===A: huge matrix ( 106 ×106 )

x0 : initial guess

mn

mnn

mn

mnn

mmnn

vvv

vAvAvA

vvvAxA

mλαλαλα

ααα

ααα

+++=

+++=

+++=

L

L

L

22211

2211

22110

1

)(

=

M

econvergenci ⇒<1λ

• If A has full rank, its eigenvectors form a basis for Rm

• Iteration

Convergence, number of iterations, it’s all about eigenvalues…

Page 25: Eigenvalues in a Nutshell

Example 2: ITERATIVE METHODS

Page 26: Eigenvalues in a Nutshell

Example 3: Dynamical Systems

( Eigenvalues don’t have the main role here, but, who are you going to complain to?)

Page 27: Eigenvalues in a Nutshell

Arnold’s Cat

• Poincare recurrence theorem:

“ A system having a finite amount of energy and confined to a finite spatial volume will, after a sufficiently long time, return to an arbitrarily small neighborhood of its initial state.”

• Vladimir I. Arnold, Russian mathematician.

⎟⎟⎠

⎞⎜⎜⎝

⎛=

2111

A

Each pixel can be assigned to a unique pair of coordinates

(a two-dimensional vector)

Page 28: Eigenvalues in a Nutshell

⎟⎟⎠

⎞⎜⎜⎝

⎛⋅⎟⎟

⎞⎜⎜⎝

⎛=⎟⎟

⎞⎜⎜⎝

⎛=

1011

1101

2111

A(mod 1)

Page 29: Eigenvalues in a Nutshell

1 2 3 5

20 31 37 42

46 47 59 63

80797877

Page 30: Eigenvalues in a Nutshell

⎟⎟⎠

⎞⎜⎜⎝

⎛=

2111

A⎟⎟⎠

⎞⎜⎜⎝

⎛→=

85.52.

61.21λ

⎟⎟⎠

⎞⎜⎜⎝

⎛−→=

52.85.

38.02λ

1)det( =A V1

V2

Page 31: Eigenvalues in a Nutshell

More Applications

•Graph theory

•Differential Equations

•PageRank

•Physics

Page 32: Eigenvalues in a Nutshell

REFERENCES

•Chen Greif. CPSC 517 Notes, UBC/CS, Spring 2007.

•Howard Anton and Chris Rorres. Elementary Linear Algebra, Applications Version, 9th Ed. John Wiley & Sons, Inc. 2005

•Humberto Madrid de la Vega. Eigenfaces: Reconocimiento digital de facciones mediante SVD. Memorias del XXXVII Congreso SMM, 2005.

•Wikipedia: Eigenvalue, eigenvector and eigenspace.http://en.wikipedia.org/wiki/Eigenvalue