+ All Categories
Home > Documents > LMS Algorithm in a Reproducing Kernel Hilbert Space

LMS Algorithm in a Reproducing Kernel Hilbert Space

Date post: 29-Jan-2016
Category:
Upload: cadee
View: 34 times
Download: 0 times
Share this document with a friend
Description:
LMS Algorithm in a Reproducing Kernel Hilbert Space. Weifeng Liu, P. P. Pokharel, J. C. Principe Computational NeuroEngineering Laboratory, University of Florida Acknowledgment: This work was partially supported by NSF grant ECS-0300340 and ECS-0601271. Outlines. Introduction - PowerPoint PPT Presentation
Popular Tags:
37
LMS Algorithm in a Reproducing Kernel Hilbert Space Weifeng Liu, P. P. Pokharel, J. C. Principe Computational NeuroEngineering Laboratory, University of Florida Acknowledgment: This work was partially supported by NSF grant ECS-0300340 and ECS- 0601271.
Transcript
Page 1: LMS Algorithm in a Reproducing Kernel Hilbert Space

LMS Algorithm in a Reproducing Kernel Hilbert Space

Weifeng Liu, P. P. Pokharel, J. C. Principe

Computational NeuroEngineering Laboratory,

University of Florida

Acknowledgment: This work was partially supported by NSF grant ECS-0300340 and ECS-0601271.

Page 2: LMS Algorithm in a Reproducing Kernel Hilbert Space

Outlines

Introduction Least Mean Square algorithm (easy) Reproducing kernel Hilbert space (tricky) The convergence and regularization analysis

(important) Learning from error models (interesting)

Page 3: LMS Algorithm in a Reproducing Kernel Hilbert Space

Introduction

Puskal (2006) –Kernel LMS

Kivinen, Smola (2004) –Online learning with kernels (more like leaky LMS)

Moody, Platt (1990’s)—Resource allocation networks (growing and pruning)

Page 4: LMS Algorithm in a Reproducing Kernel Hilbert Space

LMS (1960, Widrow and Hoff)

Given a sequence of examples from U×R:

U: a compact set of RL. The model is assumed:

The cost function:

1 1(( , ),..., ( , ))N Nu y u y

)()( nvuwy no

n

2

1

1( ) ( ( ))

N

n ni

J w y w uN

Page 5: LMS Algorithm in a Reproducing Kernel Hilbert Space

LMS

The LMS algorithm

The weight after n iteration:

0

1

1

0

( )an n n n

an n n n

w

e y w u

w w e u

1

n an i iiw e u

(1)

(2)

Page 6: LMS Algorithm in a Reproducing Kernel Hilbert Space

Reproducing kernel Hilbert space

A continuous, symmetric, positive-definite kernel ,a mapping Φ, and an inner product

H is the closure of the span of all Φ(u). Reproducing Kernel trick The induced norm

:U U R

, H

, ( ) ( )Hf u f u

1 2 1 2( ), ( ) ( , )Hu u u u 2|| || ,H Hf f f

Page 7: LMS Algorithm in a Reproducing Kernel Hilbert Space

RKHS

Kernel trick: – An inner product in the feature space– A similarity measure you needed.

Mercer’s theorem:T

M uuuu )](),...,(),([)( 21

Page 8: LMS Algorithm in a Reproducing Kernel Hilbert Space

Common kernels

Gaussian kernel

Polynomial kernel

)||||exp(),( 2jiji uuauu

( , ) ( 1)T pi j i ju u u u

Page 9: LMS Algorithm in a Reproducing Kernel Hilbert Space

Kernel LMS

Transform the input ui to Φ(ui):

Assume Φ(ui) R∈ M

The model is assumed:

The cost function:

1 1(( ( ), ),..., ( ( ), ))N Nu y u y

( ( )) ( )on ny u v n

2

1

1( ) ( ( ( )))

N

n ni

J y uN

Page 10: LMS Algorithm in a Reproducing Kernel Hilbert Space

Kernel LMS

The KLMS algorithm

The weight after n iteration:

0

1

1

0

( ( ))

( )

an n n n

an n n n

e y u

e u

1( )

n an i ii

e u

(3)

(4)

Page 11: LMS Algorithm in a Reproducing Kernel Hilbert Space

Kernel LMS

1

1

1

1

1

1

1

( ( ))

( ), ( )

( , ),

( ( )),

( ).

n n

n ai i ni H

n ai i ni

an n n n

n an i ii

u

e u u

e u u

e y u

e u

(5)

Page 12: LMS Algorithm in a Reproducing Kernel Hilbert Space

Kernel LMS

After the learning, the input-output relation:

1

( ( ))

( , )

N

N ai ii

y u

e u u

(6)

Page 13: LMS Algorithm in a Reproducing Kernel Hilbert Space

KLMS vs. RBF

KLMS:

RBF:

α satisfy

G is the gram matrix: G(i,j)=ĸ(ui,uj) RBF needs regularization. Does KLMS need regularization?

1( ) ( , )

N ai ii

y e u u

(7)

1( , )

N

i iiy u u

(8)

G y

Page 14: LMS Algorithm in a Reproducing Kernel Hilbert Space

KLMS vs. LMS

Kernel LMS is nothing but LMS in the feature space--a very high dimensional reproducing kernel Hilbert space (M>N)

Eigen-spread is awful—does it converge?

Page 15: LMS Algorithm in a Reproducing Kernel Hilbert Space

Example: MG signal predication

Time embedding: 10.

Learn rate: 0.2 500 training data 100 test data point. Gaussian noise noise variance: .04

0 100 200 300 400 5000

0.02

0.04

0.06

0.08

0.1mse linearmse kernel

Page 16: LMS Algorithm in a Reproducing Kernel Hilbert Space

Example: MG signal predication

MSE Linear LMS

KLMS RBF (λ=0)

RBF (λ=.1)

RBF (λ=1)

RBF (λ=10)

training 0.021 0.0060 0 0.0026 0.0036 0.010

test 0.026 0.0066 0.019 0.0041 0.0050 0.014

Page 17: LMS Algorithm in a Reproducing Kernel Hilbert Space

Complexity Comparison

RBF KLMS LMS

Computation O(N3) O(N2) O(L)

Memory O(N2+N*L) O(N*L) O(L)

Page 18: LMS Algorithm in a Reproducing Kernel Hilbert Space

The asymptotic analysis on convergence—small step-size theory

Denote The correlation matrix

is singular. Assume

and

1

1 NT

x i ii

R x xN

Mii Rux )(

1 1... ... 0k k M T

xR P P

Page 19: LMS Algorithm in a Reproducing Kernel Hilbert Space

The asymptotic analysis on convergence—small step-size theory

Denote

we have1( )

Mon i ii

n P

[ ( )] (1 ) (0)ni i iE n

2 2 2min min[| ( ) | ] (1 ) (| (0) | )2 2

ni i i

i i

J JE n

Page 20: LMS Algorithm in a Reproducing Kernel Hilbert Space

The weight stays at the initial place in the 0-eigen-value directions

If

we have

0i

[ ( )] (0)i iE n

2 2[| ( ) | ] | (0) |i iE n

Page 21: LMS Algorithm in a Reproducing Kernel Hilbert Space

The 0-eigen-value directions does not affect the MSE

Denote

2 2min minmin 1 1

( ) (| (0) | )(1 )2 2

M M ni i i ii i

J JJ n J

2( ) [| ( ) | ]iJ n E y x

It does not care about the null space! It only focuses on the data space!

Page 22: LMS Algorithm in a Reproducing Kernel Hilbert Space

The minimum norm initialization

The initialization gives the minimum norm possible solution.

00

M

n i iiw P

2 2 2 2

1 1 1ˆ ˆ ˆ|| || || || || || || ||

M k M

n i i ii i i kw w w

Page 23: LMS Algorithm in a Reproducing Kernel Hilbert Space

Minimum norm solution

0 2 4-1

0

1

2

3

4

5

Page 24: LMS Algorithm in a Reproducing Kernel Hilbert Space

Learning is Ill-posed

Page 25: LMS Algorithm in a Reproducing Kernel Hilbert Space

Over-learning

Page 26: LMS Algorithm in a Reproducing Kernel Hilbert Space

Regularization Technique

Learning from finite data is ill-posed. A priori information--Smoothness is needed. The norm of the function, which indicates the

‘slope’ of the linear operator is constrained. In statistical learning theory, the norm is

associated with the confidence of uniform convergence!

Page 27: LMS Algorithm in a Reproducing Kernel Hilbert Space

Regularized RBF

The cost function:

or equivalently2 2

1

1( ) ( ( ( ))) || ||

N

n ni

J y uN

2

1

2

1( ) ( ( ( )))

subject to || ||

N

n ni

J y uN

C

Page 28: LMS Algorithm in a Reproducing Kernel Hilbert Space

KLMS as a learning algorithm

The model with The following inequalities hold

The proof…(H∞ robust + triangle inequality + matrix transformation + derivative + …)

( ) ( )on ny x v n ( )n nx u

2 1 2 2|| || || || 2 || ||a oe v

2 2|| || 2 || ||ae y

Page 29: LMS Algorithm in a Reproducing Kernel Hilbert Space

The solution of regularized RBF is

The reason of ill-posedness is the inversion of the matrix (G+λI)

The numerical analysis

1( , )

N

i iiy u u

1( )G I y

1 1|| ( ) || as 0G I

Page 30: LMS Algorithm in a Reproducing Kernel Hilbert Space

The numerical analysis

The solution of KLMS is

By the inequality we have

ae Ly

1( , )

N ai ii

y e u u

|| || 2L

Page 31: LMS Algorithm in a Reproducing Kernel Hilbert Space

Example: MG signal predication

weight KLMS RBF (λ=0)

RBF (λ=.1)

RBF (λ=1)

RBF (λ=10)

norm 0.520 4.8e+3 10.90 1.37 0.231

Page 32: LMS Algorithm in a Reproducing Kernel Hilbert Space

The conclusion

The LMS algorithm can be readily used in a RKHS to derive nonlinear algorithms.

From the machine learning view, the LMS method is a simple tool to have a regularized solution.

Page 33: LMS Algorithm in a Reproducing Kernel Hilbert Space

Demo

Page 34: LMS Algorithm in a Reproducing Kernel Hilbert Space

Demo

Page 35: LMS Algorithm in a Reproducing Kernel Hilbert Space

LMS learning model

An event happens, and a decision made. If the decision is correct, nothing happens. If an error is incurred, a correction is made

on the original model. If we do things right, everything is fine and

life goes on. If we do something wrong, lessons are drawn

and our abilities are honed.

Page 36: LMS Algorithm in a Reproducing Kernel Hilbert Space

Would we over-learn?

If the real world is attempted to be modeled mathematically, what dimension is appropriate?

Are we likely to over-learn? Are we using the LMS algorithm? What is good to remember the past? What is bad to be a perfectionist?

Page 37: LMS Algorithm in a Reproducing Kernel Hilbert Space

"If you shut your door to all errors, truth will be shut out."---Rabindranath Tagore


Recommended