Professional Documents
Culture Documents
Laopti Assignment 53
min f ( x)
x
g ( x) ≥ 0
min f ( x)
x is equivalent to
g ( x) ≥ 0
Note very specially that, the whole argument is centered around the objective
function value.
Here through several pictures, we will prove one part of the relation. That is:
min f ( x)
x
⇒ min max L( x, λ ) . It is RHS of (1)
g ( x) ≥ 0 x λ
min f ( x)
⇒ max min L( x, λ )
g ( x) ≥ 0 λ x
We will do it later.
Consider following LP
max cT x
x
Ax ≤ b
x≥0
One way is to convert into a standard format (like changing max to min and chane
type inequality etc).
It may take some time and practice to fully appreciate the above statement.
max cT x
x
Ax ≤ b ⇒ L( x, y ≥ 0, λ ≥ 0) = cT x + yT ( b − Ax ) + λ T x
x≥0
max cT x
x
Ax ≥ b ⇒ min max L( x, y ≥ 0, λ ≥ 0) = cT x + yT ( b − Ax ) + λ T x
y ,λ
x≥0
x
With respect to primal variable we find optimality condition and substitute back
into lagrangian to obtain a minimization problem in terms of dual variables.
∂L
= 0 ⇒ c − AT y + λ = 0 vector
∂x
cT x + yT ( Ax − b ) + λ T x = ( AT y − λ )T x + yT ( b − Ax ) + λ T x
min bT y
y
AT y ≥ c
y≥0
Assignment Question
1
min wT w
w,γ 2
di ( wT xi − γ ) ≥ 1 ∀i = 1: m
You are about to enter into the world of kernel methods which revolutionized
machine learning theory in 1990s .
Note that in deep learning algorithms for classification, the last block is still an
SVM classifier.
Kernel PCA, Kernel CCA Kernel ICA are powerful concepts useful in AI and data
science.