artificial immune systems razieh khamseh-ashari department of electrical and computer eng isfahan...

Artificial Immune Systems

Razieh Khamseh-Ashari

Department of Electrical and Computer Eng

Isfahan University of Technology

Supervisor: Dr. Abdolreza Mirzaei

Introduction to the Immune System


A Framework to Design Artificial

Immune Systems (AIS)

Representation Schemes

Affinity Measures

Immune Algorithms

Outline

2Artificial Immune Systems






Affinity Measures

Immune Algorithms

Outline


What is the Immune System ?

a complex system of cellular and molecular components having the primary function of distinguishing self from not self and defense against foreign organisms or substances (Dorland's Illustrated Medical Dictionary)

The immune system is a cognitive system whose primary role is to provide body maintenance (Cohen)

Immune system was evolutionary selected as a consequence of its first and primordial function to provide an ideal inter-cellular communication pathway (Stewart)


1. Immune System 2. AIS Concepts 3. Designing a Framework

Artificial Immune Systems 5

Lymphatic vessels

Lymph nodes

Thymus

Spleen

Tonsils andadenoids

Bone marrow

Appendix

Peyer’s patches

Primary lymphoidorgans

Secondary lymphoidorgans


Classical Immunity

The purpose of the immune system is

defence

Innate and acquired immunity

Innate is the first line of defense. Germ line

encoded (passed from parents) and is quite

‘static’ (but not totally static)

Adaptive (acquired). Somatic (cellular) and is

acquired by the host over the life time. Very

dynamic.

These two interact and affect each other



Multiple layers of the immune system

Phagocyte

Adaptive immune

response

Lymphocytes

Innate immune

response

Biochemical barriers

Skin

Pathogens



Innate Immunity

May take days to remove an infection, if

it fails, then the adaptive response may

take over

Macrophages and neurophils are actors

Other actors such as TLR’s and

dendritic cells (next lecture) are

essential for recognition8Artificial Immune Systems


Adaptive Immune System



Lymphocytes

Carry antigen receptors that are specific

They are produced in the bone marrow

through

B and T Cells are the main actors of the

adaptive immune system



B Cell Pattern Recognition

B cells have receptors called antibodies The immune recognition is based on the

complementarity between the binding region of the receptor and a portion of the antigen calledthe epitope.

Recognition is not just by a single antibody, but a collection of them Learn not through a single agent, but

multiple ones

Epitopes

B-cell Receptors (Ab)

Antigen

B-cell

BCR or Antibody



Processes within the Immune System (very basically)

Negative Selection

Censoring of T-cells in the thymus gland of

T-cells that recognise self• Defining normal system behavior

Clonal Selection

Proliferation and differentiation of cells

when they have recognised something• Generalise and learn

Self vs Non-Self



Clonal Selection



Clonal Selection

Foreign antigens

Proliferation(Cloning)

Differentiation

Plasma cells

Memory cellsSelection

M

M

Antibody

Self-antigen

Self-antigen

Clonal deletion(negative selection)

Clonal deletion(negative selection)



Affinity Maturation

Responses mediated by T cells improve

with experience

Mutation on receptors (hypermutation and

receptor editing)

During the clonal expansion, mutation can

lead to increased affinity, these new ones are

selected to enter a ‘pool’ of memory cells• Can also lead to bad ones and these are deleted



Immune Responses

Antigen Ag1 Antigens Ag1, Ag2

Primary Response Secondary Response

Lag

Response to Ag1

Ant

ibod

y C

once

ntra

tion

Time

Lag

Response to Ag2

Response to Ag1

...

...

Cross-Reactive Response

...

...

Antigen Ag1 + Ag3

Response to Ag1 + Ag3

Lag



Summary

Innate and adaptive immunity Focused on adaptive here

Lymphocytes Negative selection Clonal selection Immune memory and learning



Further Immunology and Modelling


What is the Immune System ?

• The are many different viewpoints

• These views are not mutually exclusive

• Lots of common ingredients

classical

networkdanger



Problems with the classical view

What happens if self changes?

What about things that are “not harmful”

The Danger model was proposed



Danger theory (Matzinger 1994)

it is not “non-self”, but “danger” that the IS recognises dangerous invaders cause cell death or

stress these cells generate “danger signal”

molecules• unlike natural cell death

these stimulate an immune response local to the danger

• to identify the “culprit”



Immune Network Theory

Idiotypic network (Jerne, 1974)

B cells co-stimulate each other

Treat each other a bit like antigens

Creates an immunological memory

1

2

3

Ag

A ctiva tionP ositive response

S uppress ionN ega tive response

A ntibody

P ara tope

Id io tope





Remarkable Immune Properties

Concepts, Scope and Applications

Brief History of AIS




Affinity Measures

Immune Algorithms

Outline


Artificial Immune Systems: A Definition

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

[De Castro and Timmis,2002]



Concepts

Specificity Diversity Clonal selection Affinity maturation Immunity memory Positive and negative selection Distributing ability Multi-layering Self-organization Anomaly detection



Scope of AIS

Pattern recognition

Fault and anomaly detection

Data analysis (classification, clustering, etc.)

Agent-based systems

Search and optimization

Machine-learning

Autonomous navigation and control

Artificial life

Security of information systems26Artificial Immune Systems


Scope of AIS

Clu

ster

ing/

clas

sifi

cat i

on

Ano

mal

y de

tect

ion

Com

p ute

r se

curi

ty

Opt

imis

atio

n

Lea

rnin

g

Bi o

i nfo

rmat

ics

I mag

e pr

oc.

Ro b

oti c

s

Co n

trol

Web

mi n

i ng

0

10

20



The Early Days:

Developed from the field of theoretical immunology in the mid

1980’s.

1990 – Bersini first use of immune algorithms to solve problems

Forrest et al – Computer Security mid 1990’s

Work by IBM on virus detection

Hunt et al, mid 1990’s – Machine learning





Remarkable Immune Properties

Concepts, Scope and Applications

Brief History of AIS




Affinity Measures

Immune Algorithms

Outline


A Framework for AIS

Algorithms

Affinity

Representation

Application

Solution

AIS

Shape-Space

Binary

Integer

Real-valued

Symbolic

[De Castro and Timmis, 2002]



Shape-Space

An antibody can recognise any antigen whose complement lies within a small surrounding region of width 𝞮(the cross-reactivity threshold)

This results in a volume v𝞮 known as the recognition region of the antibody

𝞮

v𝞮

V

S

v𝞮

v𝞮

𝞮

𝞮

[Perelson,1989]

31

Shape space(or solution space)



Choice of Representation

Assume the general case: Ab = Ab1, Ab2, ..., AbLAg = Ag1, Ag2, ..., AgL

Binary representation Matching by bits

Continuous (numeric) Real or Integer, typically Euclidian

Categorical (nominal) E.g female or male of the attribute Gender.



Representation – Shape Space

Used for modeling antibody and antigen Determine a measure to calculate affinity

Hamming shape space(binary)

if Abi != Agi: 0 otherwise (XOR operator)

Antigen

Antibody



A Framework for AIS

Algorithms

Affinity

Representation

Application

Solution

AIS Euclidean

Manhattan

Hamming



Affinities: related to distance/similarity Examples of affinity measures

Euclidean

Manhattan

Hamming

L

iii AgAbD

1

2)(

L

iii AgAbD

1

L

i

ii AgAbD

1 otherwise0

if1δwhereδ,

Affinity



Affinities in Hamming Shape-Space

(a) Hamming distance

(b) r-contiguous bits rule

XOR : Affinity: 6

0 0 1 1 0 0 1 1

1 1 1 0 1 1 0 1

1 1 0 1 1 1 1 0

XOR :

0 0 1 1 0 0 1 1

1 1 1 0 1 1 0 1

1 1 0 1 1 1 1 0

Affinity: 4



Mutation - Binary

1 0 0 0 1 1 1 0 Original string

Mutated string

Bit to be mutated

1 0 0 0 0 1 1 0

Single-point mutation

1 0 0 0 1 1 1 0

0 0 0 0 0 1 1 0

Multi-point mutation

Original string

Mutated string

Bits to be mutated

Single point mutation

Multi-point mutation37Artificial Immune Systems


Affinity Proportional Mutation

Affinity maturation is controlled Proportional to

antigenic affinity (D*) = exp(-D*) =mutation rate D*= affinity =control parameter

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

D*

5ت=ت

10ت=ت

20ت=ت



A Framework for AIS

Algorithms

Affinity

Representation

Application

Solution

AIS

Bone Marrow Models

Clonal Selection

Negative Selection

Positive Selection

Immune Network Models



The Algorithms Layer

Bone Marrow Models

Clonal Selection CLONALG

AIRS

Negative Selection

Positive Selection

Network Models AINE

aiNET



Bone Marrow Models

Gene libraries are used to create antibodies from the bone marrow

Use this idea to generate attribute strings that represent receptors

Antibody production through a random concatenation from gene libraries

An individual genome corresponds to four libraries: Library 1 Library 2 Library 3 Library 4

A1 A2 A3 A4 A5 A6 A7 A8

A3 D5 C8 B2

A3 D5 C8 B2

A3 B2 C8 D5 Expressed Ab molecule

B1 B2 B3 B4 B5 B6 B7 B8 C1 C2 C3 C4 C5 C6 C7 C8 D1 D2 D3 D4 D5 D6 D7 D8



Clonal Selection –CLONALG

1. Initialisation

2. Antigenic presentation

a. Affinity evaluation

b. Clonal selection and expansion

c. Affinity maturation

d. Metadynamics

3. Cycle 43Artificial Immune Systems


1. Initialisation





d. Metadynamics

3. Cycle

CLONALG

Create a random population of individuals (P)



1. Initialisation





d. Metadynamics

3. Cycle

CLONALG

For each antigenic pattern in the data-set S do:



1. Initialisation2. Antigenic

presentation




d. Metadynamics

3. Cycle

CLONALG

Present it to the population P and determine its affinity with each element of the population



1. Initialisation





d. Metadynamics

3. Cycle

CLONALG

Select n highest affinity elements of P

Generate clones proportional to their affinity with the antigen

(higher affinity=more clones)



1. Initialisation

2. Antigenic presentationa. Affinity evaluation



d. Metadynamics

3. Cycle

CLONALG

Mutate each clone High affinity=low mutation rate

and vice-versa Add mutated individuals to

population P Reselect best individual to be

kept as memory m of the antigen presented



CLONALG

1. Initialisation





d. Metadynamics

3. Cycle

Replace a number r of individuals with low affinity with randomly generated new ones



1. Initialisation





d. Metadynamics

3. Cycle

CLONALG

Repeat step 2 until a certain stopping criteria is met



CLONALG

nci =round((β*nh )/i)



Negative Selection Algorithms

Forrest 1994: Idea taken from the negative selection of T-cells in the thymus

Applied initially to computer security


Selfstrings (S)

Generaterandom strings

(R0)Match Detector

Set (R)

Reject

No

Yes

No

Yes

Detector Set ( R )

Strings (e.g. credit card use patterns)

Match

Non-self Detected

Developing the detector set

Using the detector set


Training ALCs with negative selection



Artificial Immune Network

Timmis and Neal,2000

The ALCs interact with each other to learn the structure of a non-

self pattern

The ALCs in a network co-stimulates and/or co-suppress each

other to adapt to the non-self pattern

The stimulation level

antigen stimulation

network stimulation

network suppression



1. Initialisation2. Antigenic presentation

a. Affinity evaluationb. Clonal selection and

expansionc. Affinity maturationd. Metadynamicse. Clonal suppression

3. Network interactions4. Network suppression5. Diversity6. Cycle

aiNET

For each antigenic pattern in the data-set S do:







aiNET

Present it to the population P and determine its affinity with each element of the population







aiNET

Select n highest affinity elements of P

Generate clones proportional to their affinity with the antigen

(higher affinity=more clones)



1. Initialisation

2. Antigenic presentationa. Affinity evaluation



d. Metadynamics

e. Clonal suppression

3. Network interactions

4. Network suppression

5. Diversity

6. Cycle

aiNET

Mutate each clone High affinity=low mutation rate

and vice-versa Select h highest affinity cells

and place into memory set



aiNET





Eliminate all memory clones whose affinity with the antigen is less than a predefined threshold







aiNET

Determine similarity between each pair of network antibodies







aiNET

Eliminate all network antibodies whose affinity is less than a pre-defined threshold







aiNET

Introduce a random number of new antibodies into P







aiNET

Repeat 2 - 5 for a pre-defined number of iterations



aiNET


aiNET



The aiNet model uses the minimal spanning tree of the formed weighted-edge graph or hierarchical agglomerative clustering to determine the structure of the network clusters in the graph.

The stopping condition of the while-loop can be one of the following

Setting an iteration counter

Setting the maximum size of the network

Testing for convergence

aiNET on Data Mining

Limited visualisation

Interpret via MST or dendrogram

Compression rate of 81%

Successfully identifies the clusters

0 0.2 0.4 0.6 0.8 1

0

0.2

0.4

0.6

0.8

11 1

1

1

1

1

1

111 1

1

1

1

1

1

1

1

1

1

1 1

1

1

1

1

1

1

1

11

1

1

1

1

1

11 1

1

1

1

1

1

11

1

1

1

1

11 1

11

1

11

1

11

1

111 1

1

1

1

1

1

1

1

1

1

1

1

1

1

11

1

1

1

11

1

1

1

1

1

1

1

1

1

11

1

1

1

1

1

1

11

1

1

1

1

1

1

11

111

1

1

1

11

1

1

1

11

11

111

1 111

1

1

1

1

11

1

1

1

11

1

1

1

1

11

1

11

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

11

1

1

11

1

1

1

11

1

1

1

11

1

1

1

1

1 1

111

1

1

1

1

11

1

1

1

1

1

1

11

1

11

111

1

1

1

1

1

1

1 11 1

1

1

11

1

1

1

1

11

1

1

1

11

1

11

11

11

1

1

1

111

1

11

1

1

1

1

11

1

1

1

1

1

1

1

11

1

1

11

1

1

11

1

1

1

1 1

11

11

1

11

1

1

1

1

1

11

1

11

1

11

1

11

1

1

1

1

22

22

2 2222222

22

2222

22

2222222

22 2

222

22

2222

22 22

22 222

2 22 2222222

2222

222222

222

2222 22222222

2

2222

2222

22

2

2 22

2

2

33 3

3

33 3

3333

3

3

33

333 33

3

3 33 3333

3

3333

333

3

333 33

3 33333

3

33

3

33

33

33

3

3

33

3

333

3

333

3 33

3333333

33

3 333

33

33

33 3

3

3 3333

3

4

4

444 44

444 4

444

44

44

4444

4

4 44444

4444 4

44

4444

44

4444

44

44

4

444

4

44

44 44

4

44

44 4 4

4

44 444

4 44

44

44

4

44444

444

4

4

4

4

4

4

4

4

4

4

555

5 5

55

5

555

555

55

555

555

55 5555

5 555

5

55

5

5555 555555 5555

55555

5

55

5555 5555 5

5

555 55555

55555

5 55

55

555 5

55555

55

5

5566 66 66 6666

77

77777777

77777

7777777777

777777777777

777777777777777

7

888888

8888888888888

8888888

888888888

8888888

88888

8888888

888

Training Patterns

Training Pattern

Result immune network



aiNET on multimodal optimisation

Initial population

Final population



Results – Multi Function

aiNET CLONALG



Reference

L. N. de Castro, J. I. Timmis, Artificial immune systems as a novel

soft computing paradigm, Soft Computing 7 (2003) 526–544

Adndries P.Engelbrecht ,Computational Intelligence An introduction

D. Dasguptaa, S.Yua, et al, Recent Advances in Artificial Immune

Systems: Models and Applications, Applied Soft Computing 11

(2011) 1574–1587.

J. Zheng ,Y. Chen, A Survey of artificial immune applications, Artif

Intell Rev (2010) 34:19–34


Thank You


Questions?


artificial immune systems razieh khamseh-ashari department of electrical and computer eng isfahan...

Documents

immune system2

immune recognition

adaptive immune system

framework slide

ais concepts

cognitive system

complex system of cellular

abdolreza mirzaei slide