Machine Learning Srihari

Linear Algebra for Machine

Learning
Sargur N. Srihari
[email protected]

1
Machine Learning Srihari

What is linear algebra?

•  Linear algebra is the branch of mathematics
concerning linear equations such as
a1x1+…..+anxn=b
–  In vector notation we say aTx=b
–  Called a linear transformation of x
•  Linear algebra is fundamental to geometry, for
defining objects such as lines, planes, rotations
Linear equation a1x1+…..+anxn=b
defines a plane in (x1,..,xn) space
Straight lines define common solutions
to equations

2
Machine Learning Srihari

Why do we need to know it?

•  Linear Algebra is used throughout engineering
–  Because it is based on continuous math rather than
discrete math
•  Computer scientists have little experience with it
•  Essential for understanding ML algorithms
–  E.g., We convert input vectors (x1,..,xn) into outputs
by a series of linear transformations
•  Here we discuss:
–  Concepts of linear algebra needed for ML
–  Omit other aspects of linear algebra
3
 Linear Algebra Topics
Machine Learning Srihari

–  Scalars, Vectors, Matrices and Tensors

–  Multiplying Matrices and Vectors
–  Identity and Inverse Matrices
–  Linear Dependence and Span
–  Norms
–  Special kinds of matrices and vectors
–  Eigendecomposition
–  Singular value decomposition
–  The Moore Penrose pseudoinverse
–  The trace operator
–  The determinant
–  Ex: principal components analysis 4
Machine Learning Srihari

Scalar
•  Single number
–  In contrast to other objects in linear algebra,
which are usually arrays of numbers
•  Represented in lower-case italic x
–  They can be real-valued or be integers
•  E.g., let x ∈! be the slope of the line
–  Defining a real-valued scalar
•  E.g., let n ∈! be the number of units
–  Defining a natural number scalar

5
Machine Learning Srihari

Vector
•  An array of numbers arranged in order
•  Each no. identified by an index
•  Written in lower-case bold such as x
–  its elements are in italics lower case, subscripted
⎡ x ⎤
⎢ 1 ⎥
⎢ x2 ⎥
⎢ ⎥
x=⎢ ⎥
⎢ ⎥
⎢ ⎥
⎢ xn ⎥
⎢⎣ ⎥⎦
•  If each element is in R then x is in Rn
•  We can think of vectors as points in space
–  Each element gives coordinate along an axis
6
Machine Learning Srihari

Matrices
•  2-D array of numbers
–  So each element identified by two indices
•  Denoted by bold typeface A
–  Elements indicated by name in italic but not bold
•  A1,1 is the top left entry and Am,nis the bottom right entry
•  We can identify nos in vertical column j by writing : for the
horizontal coordinate
•  E.g., ⎡ A A ⎤
A= ⎢ ⎥
1,1 1,2

⎢ A2,1 A2,2 ⎥
⎣ ⎦

•  Ai: is ith row of A, A:j is jth column of A

•  If A has shape of height m and width n with
real-values then A ∈! m×n
7
Machine Learning Srihari

Tensor
•  Sometimes need an array with more than two
axes
–  E.g., an RGB color image has three axes
•  A tensor is an array of numbers arranged on a
regular grid with variable number of axes
–  See figure next
•  Denote a tensor with this bold typeface: A
•  Element (i,j,k) of tensor denoted by Ai,j,k
8
Machine Learning Srihari

Shapes of Tensors

9
Machine Learning Srihari

Transpose of a Matrix
•  An important operation on matrices
•  The transpose of a matrix A is denoted as AT
•  Defined as
(AT)i,j=Aj,i
–  The mirror image across a diagonal line
•  Called the main diagonal , running down to the right
starting from upper left corner

⎡ A A1,2 A1,3 ⎤ ⎡ A A2,1 A3,1 ⎤ ⎡ A A1,2 ⎤ ⎡ A A2,1 A3,1 ⎤

⎢ 1,1 ⎥ ⎢ 1,1 ⎥ ⎢ 1,1 ⎥ ⎢ 1,1 ⎥
A = ⎢ A2,1 A2,2 A2,3 ⎥ ⇒ A = ⎢ A1,2
T
A2,2 A3,2 ⎥ A = ⎢ A2,1 A2,2 ⎥⇒ A = ⎢ A
T
A2,2 A3,2 ⎥
⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ 1,2 ⎥
⎢ A3,1 A3,2 A3,3 ⎥ ⎢ A1,3 A2,3 A3,3 ⎥ ⎢ A3,1 A3,2 ⎥ ⎢ ⎥
⎣ ⎦ ⎣ ⎦ ⎣ ⎦ ⎣ ⎦
10
Machine Learning Srihari

Vectors as special case of matrix

•  Vectors are matrices with a single column
•  Often written in-line using transpose
x = [x1,..,xn]T
⎡ x ⎤
⎢ 1 ⎥
⎢ x2 ⎥
⎢ ⎥
⎥ ⇒ x = ⎡⎣x 1 ,x 2 ,..x n ⎤⎦
T
x=⎢
⎢ ⎥
⎢ ⎥
⎢ xn ⎥
⎢⎣ ⎥⎦

•  A scalar is a matrix with one element

a=aT
11
Machine Learning Srihari

Matrix Addition
•  We can add matrices to each other if they have
the same shape, by adding corresponding
elements
–  If A and B have same shape (height m, width n)
C = A+B ⇒C i,j = Ai,j +Bi,j

•  A scalar can be added to a matrix or multiplied

by a scalar D = aB+c ⇒ D = aB +c i,j i,j

•  Less conventional notation used in ML:

–  Vector added to matrix C = A+b ⇒C i,j = Ai,j +bj

•  Called broadcasting since vector b added to each row of A
12
Machine Learning Srihari

Multiplying Matrices
•  For product C=AB to be defined, A has to have
the same no. of columns as the no. of rows of B
–  If A is of shape mxn and B is of shape nxp then
matrix product C is of shape mxp

C = AB ⇒C i,j = ∑ Ai,k Bk,j

k

–  Note that the standard product of two matrices is

not just the product of two individual elements
•  Such a product does exist and is called the element-wise
product or the Hadamard product A¤B
13
Machine Learning Srihari

Multiplying Vectors
•  Dot product between two vectors x and y of
same dimensionality is the matrix product xTy
•  We can think of matrix product C=AB as
computing Cij the dot product of row i of A and
column j of B

14
Machine Learning Srihari

Matrix Product Properties

•  Distributivity over addition: A(B+C)=AB+AC
•  Associativity: A(BC)=(AB)C
•  Not commutative: AB=BA is not always true
•  Dot product between vectors is commutative:
xTy=yTx
•  Transpose of a matrix product has a simple
form: (AB)T=BTAT

15
Machine Learning Srihari

Example flow of tensors in ML

A linear classifier y= WxT+b

Vector x is converted
into vector y by
multiplying x by a matrix W

A linear classifier with bias eliminated y= WxT

Machine Learning Srihari

Linear Transformation
•  Ax=b
–  where A ∈! n×n and b ∈! n
–  More explicitly A x + A x +....+ A x = b 11 1 12 2 1n n 1
n equations in
A21 x1 + A22 x2 +....+ A2n xn = b2
n unknowns
An1 x1 + Am2 x2 +....+ An,n xn = bn

⎡ A ! A1,n ⎤ ⎡ x ⎤ ⎡ b ⎤
⎢ 1,1 ⎥ ⎢ 1 ⎥ ⎢ 1 ⎥ Can view A as a linear transformation
A= ⎢ " " " ⎥ x = ⎢ " ⎥ b= ⎢ " ⎥
⎢ A ⎥ ⎢ x ⎥ ⎢ b ⎥ of vector x to vector b
⎢ n,1 ! Ann ⎥ ⎢⎣ n ⎥⎦ ⎢⎣ n ⎥⎦
⎣ ⎦

nxn nx1 n x1

•  Sometimes we wish to solve for the unknowns

x ={x1,..,xn} when A and b provide constraints
17
Machine Learning Srihari

Identity and Inverse Matrices

•  Matrix inversion is a powerful tool to analytically
solve Ax=b
•  Needs concept of Identity matrix
•  Identity matrix does not change value of vector
when we multiply the vector by identity matrix
–  Denote identity matrix that preserves n-dimensional
vectors as In
–  Formally In ∈! n×n and ∀x ∈! n ,Inx = x
⎡ 1 0 0 ⎤

–  Example of I3 ⎢
0 1 0
⎥
⎢ ⎥
⎢ 0 0 1 ⎥ 18
⎣ ⎦
Machine Learning Srihari

Matrix Inverse
•  Inverse of square matrix A defined as
A−1 A = In
•  We can now solve Ax=b as follows:
Ax = b
A−1 Ax = A−1b
In x = A−1b
−1
x = A b

•  This depends on being able to find A-1

•  If A-1 exists there are several methods for
finding it
19
Machine Learning Srihari

Solving Simultaneous equations

•  Ax = b
where A is (M+1) x (M+1)
x is (M+1) x 1: set of weights to be determined
b is N x 1

20
 Example: System of Linear
Machine Learning Srihari

Equations in Linear Regression

•  Instead of Ax=b
•  We have Φw = t
–  where Φ is m x n design matrix of m features for n
samples xj, j=1,..n
–  w is weight vector of m values
–  t is target values of sample, t=[t1,..tn]
–  We need weight w to be used with m features to
determine output

m
y(x,w)=∑wi x i
i=1 21
Machine Learning Srihari

Closed-form solutions
•  Two closed-form solutions
1. Matrix inversion x=A-1b
2. Gaussian elimination

22
Machine Learning Srihari

Linear Equations: Closed-Form Solutions

1. Matrix Formulation: Ax=b

Solution: x=A-1b

2. Gaussian Elimination
followed by back-substitution

L2-3L1àL2 L3-2L1àL3 -L2/4àL2

Machine Learning Srihari

Disadvantage of closed-form solutions

•  If A-1 exists, the same A-1 can be used for any
given b
–  But A-1 cannot be represented with sufficient
precision
–  It is not used in practice
•  Gaussian elimination also has disadvantages
–  numerical instability (division by small no.)
–  O(n3) for n x n matrix
•  Software solutions use value of b in finding x
–  E.g., difference (derivative) between b and output is
used iteratively 24
Machine Learning Srihari

How many solutions for Ax=b exist?

•  System of equations with A11 x1 + A12 x2 +....+ A1n xn = b1
A21 x1 + A22 x2 +....+ A2n xn = b2
–  n variables and m equations is:
Am1 x1 + Am2 x2 +....+ Amn xn = bm

•  Solution is x=A-1b
•  In order for A-1 to exist Ax=b must have
exactly one solution for every value of b
–  It is also possible for the system of equations to
have no solutions or an infinite no. of solutions for
some values of b
•  It is not possible to have more than one but fewer than
infinitely many solutions
–  If x and y are solutions then z=α x + (1-α) y is a
solution for any real α 25
Machine Learning Srihari

Span of a set of vectors

•  Span of a set of vectors: set of points obtained
by a linear combination of those vectors
–  A linear combination of vectors {v(1),.., v(n)} with
coefficients ci is ∑ c v i
(i)

–  System of equations is Ax=b

•  A column of A, i.e., A:i specifies travel in direction i
•  How much we need to travel is given by xi
•  This is a linear combination of vectors Ax = ∑ x i A:, i
i

–  Thus determining whether Ax=b has a solution is

equivalent to determining whether b is in the span of
columns of A
•  This span is referred to as column space or range of A
Machine Learning Srihari

Conditions for a solution to Ax=b

•  Matrix must be square, i.e., m=n and all
columns must be linearly independent
–  Necessary condition is n ≥ m
•  For a solution to exist when A ∈!
m×n
we require the
column space be all of ! m
–  Sufficient Condition
•  If columns are linear combinations of other columns,
m
column space is less than !
–  Columns are linearly dependent or matrix is singular
•  For column space to encompass ! m at least one set
of m linearly independent columns
•  For non-square and singular matrices
–  Methods other than matrix inversion are used
Machine Learning Srihari

Use of a Vector in Regression

•  A design matrix
–  N samples, D features

•  Feature vector has three dimensions

•  This is a regression problem
28
Machine Learning Srihari

Norms
•  Used for measuring the size of a vector
•  Norms map vectors to non-negative values
•  Norm of vector x = [x1,..,xn]T is distance from
origin to x
–  It is any function f that satisfies:
( )
f x = 0 ⇒x = 0
( ) ( ) Triangle Inequality
f( x +y) ≤ f x + f y
∀α ∈R f (αx ) = α f (x )

29
 LP Norm
Machine Learning Srihari

•  Definition: 1
⎛ p⎞ p
x = ⎜ ∑ xi ⎟
p ⎝ i ⎠
–  L2 Norm
•  Called Euclidean norm
–  Simply the Euclidean distance
between the origin and the point x
–  written simply as ||x||
–  Squared Euclidean norm is same as xTx 22 + 22 = 8 = 2 2

–  L1 Norm
•  Useful when 0 and non-zero have to be distinguished
–  Note that L2 increases slowly near origin, e.g., 0.12=0.01)
–  L∞ Norm x = max x i
∞ i

•  Called max norm 30

Machine Learning Srihari

Use of norm in Regression

•  Linear Regression
x: a vector, w: weight vector
y(x,w) = w0+w1x1+..+wd xd = wTx

With nonlinear basis functions ϕj

M −1
y(x,w) = w0 + ∑ w φ (x)
j j
j =1

•  Loss Function
1 N
λ
E(w) = ∑ {y(x n ,w) − tn } + || w 2 ||
! 2

2 n =1 2

Second term is a weighted norm

called a regularizer (to prevent overfitting) 31
Machine Learning Srihari

LP Norm and Distance

•  Norm is the length of a vector

•  We can use it to draw a unit circle from origin

–  Different P values yield different shapes
•  Euclidean norm yields a circle

•  Distance between two vectors (v,w)

–  dist(v,w)=||v-w||
= (v − w )
1 1
2
+ .. + (vn − wn )2
Distance to origin would just be sqrt of sum of squares 32
Machine Learning Srihari

Size of a Matrix: Frobenius Norm

•  Similar to L2 norm ⎛ ⎞
A = ⎜ ∑ Ai,j2 ⎟
1
2

F
⎝ i,j ⎠

⎡ 2 −1 5 ⎤
⎢ ⎥
A=⎢ 0 2 1 ⎥ A = 4 + 1 + 25 + .. + 1 = 46
⎢ 3 1 1 ⎥ V matrix W matrix
⎣ ⎦

I J K
•  Frobenius in ML I1×(I+1) × V(I+1)×J=netJ
–  Layers of neural network hj=f(netj) f(x)=1/(1+e-x)
involve matrix multiplication
–  Regularization:
•  minimize Frobenius of weight
matrices ||W(i)|| over L layers

33
Machine Learning Srihari

Angle between Vectors

•  Dot product of two vectors can be written in
terms of their L2 norms and angle θ between
them T
x y ⇒||x ||2||y ||2 cos θ

•  Cosine between two vectors is a measure of

their similarity

34
Machine Learning Srihari

Special kind of Matrix: Diagonal

•  Diagonal Matrix has mostly zeros, with non-
zero entries only in diagonal
–  E.g., identity matrix, where all diagonal entries are 1

–  E.g., covariance matrix with independent features

⎧⎪ 1 ⎫
N(x | µ,Σ) =
1 1
exp ⎪⎨− (x − µ)T Σ−1(x − µ)⎪⎪⎬
(2π)D/2 | Σ |1/2 ⎪⎪⎩ 2 ⎪⎪⎭
If Cov(X,Y)=0 then E(XY)=E(X)E(Y)
Machine Learning Srihari

Efficiency of Diagonal Matrix

•  diag (v) denotes a square diagonal matrix with
diagonal elements given by entries of vector v
•  Multiplying vector x by a diagonal matrix is
efficient
–  To compute diag(v)x we only need to scale each xi
by vi
diag( v)x = v ⊙x
•  Inverting a square diagonal matrix is efficient
–  Inverse exists iff every diagonal entry is nonzero, in
which case diag (v)-1=diag ([1/v1,..,1/vn]T)
Machine Learning Srihari

Special kind of Matrix: Symmetric

•  A symmetric matrix equals its transpose: A=AT
–  E.g., a distance matrix is symmetric with Aij=Aji

–  E.g., covariance matrices are symmetric

Machine Learning Srihari

Special Kinds of Vectors

•  Unit Vector
–  A vector with unit norm x =1
2

•  Orthogonal Vectors
–  A vector x and a vector y are
orthogonal to each other if xTy=0
•  If vectors have nonzero norm, vectors at
90 degrees to each other
–  Orthonormal Vectors
•  Vectors are orthogonal & have unit norm
•  Orthogonal Matrix
–  A square matrix whose rows are mutually
orthonormal: ATA=AAT=I Orthogonal matrices are of
–  A-1=AT interest because their inverse is
very cheap to compute
Machine Learning Srihari

Matrix decomposition
•  Matrices can be decomposed into factors to
learn universal properties, just like integers:
–  Properties not discernible from their representation
1. Decomposition of integer into prime factors
•  From 12=2×2×3 we can discern that
–  12 is not divisible by 5 or
–  any multiple of 12 is divisible by 3
–  But representations of 12 in binary or decimal are different
2. Decomposition of Matrix A as A=Vdiag(λ)V-1
•  where V is formed of eigenvectors and λ are eigenvalues,
e.g,
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
A= ⎢ 2 1 ⎥ has eigenvalues λ=1 and λ=3 and eigenvectors V: vλ=1 = ⎢ 1 ⎥ ,vλ=3 = ⎢ 1 ⎥
⎢ −1 ⎥ ⎢ 1 ⎥
⎢ 1 2 ⎥ ⎣ ⎦ ⎣ ⎦
⎣ ⎦
Machine Learning Srihari

Eigenvector
•  An eigenvector of a square matrix
A is a non-zero vector v such that
multiplication by A only changes
the scale of v
Av=λv
–  The scalar λ is known as eigenvalue
•  If v is an eigenvector of A, so is
any rescaled vector sv. Moreover Wikipedia

sv still has the same eigen value.

Thus look for a unit eigenvector
40
Machine Learning Srihari

Eigenvalue and Characteristic Polynomial

⎡ A L A1,n ⎤ ⎡ v ⎤ ⎡ w ⎤
⎢ 1,1 ⎥ ⎢ 1 ⎥ ⎢ 1 ⎥

•  Consider Av=w A= ⎢ M M
⎢
⎢ An,1 L
⎣
M ⎥
Ann ⎥
⎥
⎦
v= ⎢ M
⎢ v
⎢⎣ n
⎥
⎥
⎥⎦
w= ⎢ M ⎥
⎢ w ⎥
⎢⎣ n ⎥⎦

•  If v and w are scalar multiples, i.e., if Av=λv

•  then v is an eigenvector of the linear transformation A
and the scale factor λ is the eigenvalue corresponding
to the eigen vector
•  This is the eigenvalue equation of matrix A
–  Stated equivalently as (A-λI)v=0
–  This has a non-zero solution if |A-λI|=0 as
•  The polynomial of degree n can be factored as
|A-λI| = (λ1-λ)(λ2-λ)…(λn-λ)
•  The λ1, λ2…λn are roots of the polynomial and are
eigenvalues of A
Machine Learning Srihari

Example of Eigenvalue/Eigenvector
•  Consider the matrix ⎡ 2 1 ⎤
A= ⎢ ⎥
⎢ 1 2 ⎥
⎣ ⎦
•  Taking determinant of (A-λI), the char poly is
⎡ 2 −λ 1 ⎤
| A −λI |= ⎢⎢ ⎥ = 3 − 4λ + λ 2
⎥
⎣ 1 2 −λ ⎦

•  It has roots λ=1 and λ=3 which are the two

eigenvalues of A
•  The eigenvectors are found by solving for v in
Av=λv, which are ⎡ 1 ⎤ ⎡ 1 ⎤
vλ=1 = ⎢ ⎥ ,v = ⎢ ⎥
⎢ −1 ⎥ λ=3 ⎢ 1 ⎥
⎣ ⎦ ⎣ ⎦ 42
Machine Learning Srihari

Eigendecomposition
•  Suppose that matrix A has n linearly
independent eigenvectors {v(1),..,v(n)} with
eigenvalues {λ1,..,λn}
•  Concatenate eigenvectors to form matrix V
•  Concatenate eigenvalues to form vector
λ=[λ1,..,λn]
•  Eigendecomposition of A is given by
A=Vdiag(λ)V-1
43
Machine Learning Srihari

Decomposition of Symmetric Matrix

•  Every real symmetric matrix A can be
decomposed into real-valued eigenvectors and
eigenvalues
A=QΛQT
where Q is an orthogonal matrix composed of
eigenvectors of A: {v(1),..,v(n)}
orthogonal matrix: components are orthogonal or v(i)Tv(j)=0
Λ is a diagonal matrix of eigenvalues {λ1,..,λn}
•  We can think of A as scaling space by λi in
direction v(i)
–  See figure on next slide 44
Machine Learning Srihari

Effect of Eigenvectors and Eigenvalues

•  Example of 2×2 matrix
•  Matrix A with two orthonormal eigenvectors
–  v(1) with eigenvalue λ1, v(2) with eigenvalue λ2
Plot of unit vectors u ∈!
2
Plot of vectors Au
(circle) (ellipse)

with two variables x1 and x2

45
Machine Learning Srihari

Eigendecomposition is not unique

•  Eigendecomposition is A=QΛQT
–  where Q is an orthogonal matrix composed of
eigenvectors of A
•  Decomposition is not unique when two
eigenvalues are the same
•  By convention order entries of Λ in descending
order:
–  Under this convention, eigendecomposition is
unique if all eigenvalues are unique
46
Machine Learning Srihari

What does eigendecomposition tell us?

•  Tells us useful facts about the matrix:
1.  Matrix is singular if & only if any eigenvalue is zero
2.  Useful to optimize quadratic expressions of form
f(x)=xTAx subject to ||x||2=1
Whenever x is equal to an eigenvector, f is equal to the
corresponding eigenvalue
Maximum value of f is max eigen value, minimum value is
min eigen value
Example of such a quadratic form appears in multivariate
Gaussian ⎧
⎪ ⎫
⎪
1 1 ⎪ 1
N(x | µ,Σ) = exp ⎨− (x − µ)T Σ−1(x − µ)⎪
⎬
D/2
(2π) | Σ | 1/2 ⎪
⎪
⎩ 2 ⎪
⎪
⎭
47
Machine Learning Srihari

Positive Definite Matrix

•  A matrix whose eigenvalues are all positive is
called positive definite
–  Positive or zero is called positive semidefinite
•  If eigen values are all negative it is negative
definite
–  Positive definite matrices guarantee that xTAx ≥ 0

48
Machine Learning Srihari

Singular Value Decomposition (SVD)

•  Eigendecomposition has form: A=Vdiag(λ)V-1
–  If A is not square, eigendecomposition is undefined
•  SVD is a decomposition of the form A=UDVT
•  SVD is more general than eigendecomposition
–  Used with any matrix rather than symmetric ones
–  Every real matrix has a SVD
•  Same is not true of eigen decomposition
Machine Learning Srihari

SVD Definition
•  Write A as a product of 3 matrices: A=UDVT
–  If A is m×n, then U is m×m, D is m×n, V is n×n
•  Each of these matrices have a special structure
•  U and V are orthogonal matrices
•  D is a diagonal matrix not necessarily square
–  Elements of Diagonal of D are called singular values of A
–  Columns of U are called left singular vectors
–  Columns of V are called right singular vectors

•  SVD interpreted in terms of eigendecomposition

•  Left singular vectors of A are eigenvectors of AAT
•  Right singular vectors of A are eigenvectors of ATA
•  Nonzero singular values of A are square roots of eigen
values of ATA. Same is true of AAT
Machine Learning Srihari

Use of SVD in ML
1.  SVD is used in generalizing matrix inversion
–  Moore-Penrose inverse (discussed next)
2.  Used in Recommendation systems
–  Collaborative filtering (CF)
•  Method to predict a rating for a user-item pair based on the
history of ratings given by the user and given to the item
•  Most CF algorithms are based on user-item rating matrix
where each row represents a user, each column an item
–  Entries of this matrix are ratings given by users to items
•  SVD reduces no.of features of a data set by reducing space
dimensions from N to K where K < N
51
Machine Learning Srihari

SVD in Collaborative Filtering

•  X is the utility matrix

–  Xij denotes how user i likes item j
–  CF fills blank (cell) in utility matrix that has no entry
•  Scalability and sparsity is handled using SVD
–  SVD decreases dimension of utility matrix by
extracting its latent factors
•  Map each user and item into latent space of dimension r
52
Machine Learning Srihari

Moore-Penrose Pseudoinverse
•  Most useful feature of SVD is that it can be
used to generalize matrix inversion to non-
square matrices
•  Practical algorithms for computing the
pseudoinverse of A are based on SVD
A+=VD+UT
–  where U,D,V are the SVD of A
•  Pseudoinverse D+ of D is obtained by taking the
reciprocal of its nonzero elements when taking transpose
of resulting matrix
53
Machine Learning Srihari

Trace of a Matrix
•  Trace operator gives the sum of the elements
along the diagonal
Tr(A )= ∑ Ai ,i
i ,i

•  Frobenius norm of a matrix can be represented

as
1

F
(
A = Tr(A)) 2

54
Machine Learning Srihari

Determinant of a Matrix
•  Determinant of a square matrix det(A) is a
mapping to a scalar
•  It is equal to the product of all eigenvalues of
the matrix
•  Measures how much multiplication by the
matrix expands or contracts space

55
Machine Learning Srihari

Example: PCA
•  A simple ML algorithm is Principal Components
Analysis
•  It can be derived using only knowledge of basic
linear algebra

56
Machine Learning Srihari

PCA Problem Statement

•  Given a collection of m points {x(1),..,x(m)} in
Rn represent them in a lower dimension.
–  For each point x(i) find a code vector c(i) in Rl
–  If l is smaller than n it will take less memory to
store the points
–  This is lossy compression
–  Find encoding function f (x) = c and a decoding
function x ≈ g ( f (x) )

57
Machine Learning Srihari

PCA using Matrix multiplication

•  One choice of decoding function is to use
matrix multiplication: g(c) =Dc where D ∈! n×l

–  D is a matrix with l columns

•  To keep encoding easy, we require columns of
D to be orthogonal to each other
–  To constrain solutions we require columns of D to
have unit norm
•  We need to find optimal code c* given D
•  Then we need optimal D
58
Machine Learning Srihari

Finding optimal code given D

•  To generate optimal code point c* given input
x, minimize the distance between input point x
and its reconstruction g(c*)
c* = argmin x − g(c)
2
c

–  Using squared L2 instead of L2, function being

minimized is equivalent to
(x
− g(c))T
(x − g(c))
•  Using g(c)=Dc optimal code can be shown to
be equivalent to c* = argmin − 2x T Dc+cT c
c 59
Machine Learning Srihari

Optimal Encoding for PCA

∇c (−2x T Dc+c T c) = 0
•  Using vector calculus
−2DT x+2c = 0
c = D x
T

•  Thus we can encode x using a matrix-vector

operation
–  To encode we use f(x)=DTx
–  For PCA reconstruction, since g(c)=Dc we use
r(x)=g(f(x))=DDTx
–  Next we need to choose the encoding matrix D

60
Machine Learning Srihari

Method for finding optimal D

•  Revisit idea of minimizing L2 distance between
inputs and reconstructions
–  But cannot consider points in isolation
–  So minimize error over all points: Frobenius norm
1
⎛
( ( ))
⎞
2 2
D* = argmin ⎜ ∑ x (ij ) − r x (i ) ⎟
D ⎝ i,j j ⎠

•  subject to DTD=Il
•  Use design matrix X, X ∈!
m×n

–  Given by stacking all vectors describing the points

•  To derive algorithm for finding D* start by
considering the case l =1
61
–  In this case D is just a single vector d
Machine Learning Srihari

Final Solution to PCA

•  For l =1, the optimization problem is solved
using eigendecomposition
–  Specifically the optimal d is given by the
eigenvector of XTX corresponding to the largest
eigenvalue
•  More generally, matrix D is given by the l
eigenvectors of X corresponding to the largest
eigenvalues (Proof by induction)

62