Docstoc

Artificial Immune Systems - An Overview

Document Sample
Artificial Immune Systems - An Overview Powered By Docstoc
					Artificial Immune Systems: An Overview
Jon Timmis
Computing Laboratory
University of Kent at Canterbury CT2 7NF. UK. J.Timmis@kent.ac.uk http:/www.cs.kent.ac.uk/~jt6

AIS – March 2004 – 1

What do I want to achieve?
l

Give you a taster of what AIS is all about


Why do we find the immune system useful?


  l

Explain what AIS are
Show you where they are being used Comments for the future Talk about all areas of AIS and applications Talk too much about how AIS relate to other bioinspired ideas (although I will mention it) Go into too much detail: this is an introduction

I won’t:
 



AIS – March 2004 – 2

Outline
lWhat lUseful

are AIS? immunology

lThinking lThe

about AIS
Areas

lApplication

Future

AIS – March 2004 – 3

Why the Immune System?
l

Recognition
 

Anomaly detection Noise tolerance

l
l l l l l l

Robustness
Changing nature of self Diversity, Adaptive Reinforcement learning Memory; Dynamically changing coverage Distributed Multi-layered

AIS – March 2004 – 4

A Definition
AIS are adaptive systems inspired by theoretical immunology and observed immune functions, principles and models, which are applied to complex problem domains

AIS – March 2004 – 5

Some History
l

Developed from the field of theoretical immunology in the mid 1980’s.
 Suggested

we ‘might look’ at the IS

l

1990 – Bersini first use of immune algos to solve problems Forrest et al – Computer Security mid 1990’s

l

l

Hunt et al, mid 1990’s – Machine learning

AIS – March 2004 – 6

Scope of AIS:
l

Computer Security(Forrest’94’96’98, Kephart’94,

l l l

l

l

Lamont’98’01,02, Dasgupta’99’01, Bentley’00’01,02) Anomaly Detection (Dasgupta’96’01’02) Fault Diagnosis (Ishida’92’93, Ishiguro’94) Data Mining & Retrieval (Hunt’95’96, Timmis’99’ 01, ’02) Pattern Recognition (Forrest’93, Gibert’94, de Castro ’02) Adaptive Control (Bersini’91)
AIS – March 2004 – 7

Scope of AIS (Cont……):
l l l l l

l
l l

Job shop Scheduling (Hart’98, ’01, ’02) Chemical Pattern Recognition (Dasgupta’99) Robotics (Ishiguro’96’97,Singh’01) Optimization (DeCastro’99,Endo’98, de Castro ’02) Web Mining (Nasaroui’02) Fault Tolerance (Tyrrell, ’01, ’02, Timmis ’02) Autonomous Systems (Varela’92,Ishiguro’96) Engineering Design Optimization (Hajela’96 ’98,

Nunes’00)

l

And so on …
AIS – March 2004 – 8

Outline
lWhat

are AIS? immunology

lUseful

lThinking lThe

about AIS
Areas

lApplication

Future

AIS – March 2004 – 9

How does it work: A simplistic view

AIS – March 2004 – 10

Learning

Primary Response
Antibody Concentration

Secondary Response

Cross-Reactive Response

Lag Lag

Lag

Response to Ag1 Response to Ag1

...
Response to Ag2

Response to Ag1 + Ag3

... ...
Antigen Ag1 Antigens Ag1, Ag2

...
Antigen Ag1 + Ag3 Time

AIS – March 2004 – 11

Immune Network Theory
lIdiotypic lB

network (Jerne, 1974)

cells co-stimulate each other


Treat each other a bit like antigens
an immunological memory

lCreates

Paratope

Suppression Negative response Ag 1 Idiotope 2 3

Antibody Activation Positive response
AIS – March 2004 – 12

Danger Theory
Proposed by Polly Matzinger, around 1995 Problem: Traditional self/non-self theory doesn’t always match observations


l l

Immune system always responds to non-self
…apart from the nonself it doesn’t respond to (harmless foreign)



Immune system always tolerates self
…apart from the self it doesn’t tolerate (dangerous self)

l

T-cell activation by APCs

AIS – March 2004 – 13

Immune System: Summary
l

The host has to distinguish either between self/non-self or dangerous/non-dangerous When an entity is recognised as foreign (or dangerous)- activate several defense mechanisms leading to its destruction (or neutralization). Subsequent exposure to similar entity results in rapid immune response. Overall behavior of the immune system is an emergent property of many local interactions. So it is useful?

l

l

l

l

AIS – March 2004 – 14

Outline
lWhat

are AIS? immunology

lUseful

lThinking lThe

about AIS
Areas

lApplication

Future

AIS – March 2004 – 15

AIS ..
l

Think about the use of AIS in terms of:
 Representation;  Affinity  Immune

Measures;
Algorithms;

l l

…and how we apply them and to what we apply them Think about the bias of all of the above and what affect that may have on the result (if any).

AIS – March 2004 – 16

Basic Immune Algorithms
lPopulation   

based:

Negative Selection Algorithms
Clonal Selection Algorithm Bone Marrow Algorithms based Immune Network Algorithms

lNetwork 

AIS – March 2004 – 17

Outline
lWhat

are AIS? immunology

lUseful

lThinking lThe

about AIS
Areas

lApplication

Future

AIS – March 2004 – 18

Data mining
lMore

benchmark problem in this case a set of labelled vectors

lAssume

lClassification

AIS – March 2004 – 19

AIRS: (Artificial Immune Recognition System)
lClonal

Selection of cells: hypermutation binding

lSurvival  lSomatic

Resource allocation

lAntibody/antigen

AIS – March 2004 – 20

AIRS: Mapping from IS to AIS
Antibody Recognition Feature Vector, Class Combination of feature vector and vector class

Antigens
Immune Memory

Training Data
Memory cells—set of mutated ARBs

AIS – March 2004 – 21

AIRS Algorithm
l l l

Data normalisation and initialisation

Memory cell identification and ARB generation
Competition for resources in the development of a candidate memory cell Potential introduction of the candidate memory cell into the set of established memory cells

l

AIS – March 2004 – 22

Classification Accuracy
lImportant

to maintain accuracy

AIRS1: Accuracy

AIRS2: Accuracy

Iris

96.7

96.0

Ionosphere

94.9

95.6

Diabetes

74.1

74.2

Sonar

84.0

84.9

AIS – March 2004 – 23

Features
lNo

need to know best architecture to get good results settings within a few percent of the best it can

lDefault

get

parameters optimize performance for a given problem set
lGeneralization

lUser-adjustable

and data reduction

AIS – March 2004 – 24

Unsupervised Learning
lAgain, lWe  

a benchmark problem in this case a set of unlabelled vectors

lAssume

can ask the questions:
Is there a large amount of redundancy? Are there any groups or subgroups intrinsic to the data? What is the structural or spatial distribution?



AIS – March 2004 – 25

MARIA: Immune principles employed
lB-cells

(antibodies)

lAntigens lAntibody/antigen lClonal lMulti-layered

binding

selection process system

AIS – March 2004 – 26

Data mining: Clustering (MARIA)
l l l l

Limited visualisation (3-d only) Good identification of clusters

High compression rate
Use vector quantisation to assess classification accuracy


Comparable with SOFM, but less parameters and no need to define structure

AIS – March 2004 – 27

Fun Immune Inspired Ideas at Kent
l

l

Immunised Fault Tolerance for Mechatronic devices  Exploratory industrial sponsored work  Increase availability of machines  Prediction of machine states  Network wide immune system that is adaptable and can disseminate (or vaccinate) other devices Immune Algorithms for Immunoinformatics  Prediction of the binding of T-cells and proteins  Very hard problem  Using AIRS to do this, so far, quite good.  Will try and exploit the nature of the problem to solve it!

AIS – March 2004 – 28

More fun …
l

l

Danger theory and Web content mining  Extracting information from the content of web pages.  Little AIS research in this area.  A very large and very dynamic data set. Domain characteristics include:  Highly volatile data.  High volume of data.  Ever-changing content.  Need for continuous adaptation.
AIS – March 2004 – 29

Mutation Testing (Motivation)
lVariability   

in software engineering:

Programming languages

Paradigms
Programming teams needs testing

lSoftware lTest

suites wear out

AIS – March 2004 – 30

Mutation Testing Fundamentals
Programmer Hypothesis:  Competent programmers make small coding errors lCoupling Effect:  Large coding errors are combinations of smaller errors a program P is incorrect, the correct version differs by only a small amount lFor all mutations M of P, find test data that either:  Causes P to behave incorrectly  Distinguishes M from P
lIf lCompetent

AIS – March 2004 – 31

Problems with Traditional Mutation Testing
lComputationally 

expensive:

Mutants
 

Large numbers Traditionally interpretive execution

lDifficult
 

to automate completely:
Generating test data
Detecting equivalent mutants

AIS – March 2004 – 32

General Aims
studying the operation of a programming team over time:
 lBy

Identify program mutation operators that frequently produce live mutants


Live mutants may reveal common errors



Derive more discriminating test data

AIS – March 2004 – 33

The “Vaccine”
set of test cases that kill the largest proportion of mutants
 lSmallest

More discriminating data

set of mutation operators that deliver the largest set of consistently living mutants


lSmallest

Overcoming common errors

AIS – March 2004 – 34

Web Mining (AISEC)
l l l

A study of performance and characteristics of an AIS applied to text mining in a dynamic domain User behaviour and interaction with e-mail can be similar to web pages Supervised classification algorithm  E-mail classified as interesting and uninteresting  Uses constant(ish) feedback from user  Capable of continuous adaptation  This tracks concept drift and can also handle A specialised AIS algorithm based in part on the immune principle of clonal selection  No previously documented algorithm was suited for use in this situation without extensive changes
AIS – March 2004 – 35

concept shift

l

AISEC (2)
l

AISEC has produced promising results and appears robust  Interesting note: Typical accuracy similar to published results from other AIS for text classification (both traditional and continuous learning)  Use a larger training set and optimise (the many) parameters  Detect when there are the optimum number of cells

l

Future work  We are researching adaptive systems for retrieval of interesting information, not necessarily purely accurate information  AISEC has been useful providing some evidence AIS applied to this domain would be possible

AIS – March 2004 – 36

Outline
lWhat

are AIS? immunology

lUseful

lThinking lThe

about AIS
Areas

lApplication

Future

AIS – March 2004 – 37

The Future
l l

l l l

Rapidly emerging field Much work is very diverse  More formal approach required? Wide possible application domains What is it that makes the immune system unique? More work with immunologists  Theories such as Danger theory, Self-Assertion may have something to say to AIS

AIS – March 2004 – 38

Integrating: Homeostasis
Ability of an organism to achieve a steady state of internal body function in a varying environment
l

Lots of complex interactions
  

Nervous system Endocrine system Immune System

l

Developed a simple neural network and endocrine controller for a mobile robot

AIS – March 2004 – 39

The Future (2)
l

l

l l l

ARTIST: A Network for Artificial Immune Systems (EPSRC funded network)  http://www.artificial-immune-systems.org Work towards:  A theoretical foundation for AIS as a new CI  Extraction of accurate metaphors  Immune System Modelling  Application of AIS Train PhD students Fund workshops/meetings Coordinate and Disseminate UK based AIS research (links to Europe)

AIS – March 2004 – 40

AIS Resources: Books
l

Artificial Immune Systems and Their Applications by Dipankar Dasgupta (Editor) Springer Verlag, January 1999. Artificial Immune Systems: A New Computational Intelligence Approach by Leandro N. de Castro, Jonathan Timmis, Springer Verlag, November 2002.

l

l

Immunocomputing: Principles and Applications by Alexander O. Tarakanov, Victor A. Skormin, Svetlana P. Sokolova, Springer Verlag, April 2003.

AIS – March 2004 – 41


				
DOCUMENT INFO
Shared By:
Categories:
Stats:
views:411
posted:4/29/2008
language:English
pages:41