1 BY: Nazanin Asadi Zohre Molaei Isfahan University of Technology.

Slides:



Advertisements
Similar presentations
Negative Selection Algorithms at GECCO /22/2005.
Advertisements

V-Detector: A Negative Selection Algorithm Zhou Ji, advised by Prof. Dasgupta Computer Science Research Day The University of Memphis March 25, 2005.
Artificial Immune Systems. CBA - Artificial Immune Systems Artificial Immune Systems: A Definition AIS are adaptive systems inspired by theoretical immunology.
CIFD: Computational Immunology for Fraud Detection
Artificial Immune Systems Razieh Khamseh-Ashari Department of Electrical and Computer Eng Isfahan University of Technology Supervisor: Dr. Abdolreza Mirzaei.
The Immune System Specific Immunity. What You Should Know Immune surveillance A range of white blood cells constantly circulate monitoring the tissues.
EvoNet Flying Circus Introduction to Evolutionary Computation Brought to you by (insert your name) The EvoNet Training Committee The EvoNet Flying Circus.
Lecture outline The nomenclature of Immunology Types of immunity (innate and adaptive; active and passive; humoral and cell- mediated) Features of immune.
Anomaly Detection in Data Docent Xiao-Zhi Gao
Lecture outline The nomenclature of Immunology
Artificial Immune Systems Andrew Watkins. Why the Immune System? Recognition –Anomaly detection –Noise tolerance Robustness Feature extraction Diversity.
Information Processing in Living Systems. Does the brain compute?
Artificial Immune Systems: An Emerging Technology Dr. Jonathan Timmis Computing Laboratory University of Kent at Canterbury England. UK.
EvoNet Flying Circus Introduction to Evolutionary Computation Brought to you by (insert your name) The EvoNet Training Committee The EvoNet Flying Circus.
Complexity and the Immune System. Why look at the immune system? -Intermediate level -One of the major information processing systems in the body (with.
Artificial Immune Systems: A New Computational Intelligence Approach
1 Negative selection algorithms: from the thymus to V-detector Dissertation defense Zhou Ji Major professor: Prof. Dasgupta Advisory committee: Dr. Lin,
Introduction to Artificial Immune Systems (AIS) BIC 2005: International Symposium on Bio-Inspired Computing Johor, MY, 5-7 September 2005 Dr. Leandro Nunes.
Artificial Immune Systems Our body’s immune system is a perfect example of a learning system. It is able to distinguish between good cells and potentially.
An Introduction to Artificial Immune Systems Dr. Jonathan Timmis Computing Laboratory University of Kent at Canterbury CT2 7NF. UK.
Dr. Fang-Ping Huang Department of Pathology, Immunology Division, Room 216 Tel
Specific immune system
Population-based metaheuristics Nature-inspired Initialize a population A new population of solutions is generated Integrate the new population into the.
By : Anas Assiri.  Introduction  fraud detection  Immune system  Artificial immune system (AIS)  AISFD  Clonal selection.
Distributed Network Intrusion Detection An Immunological Approach Steven Hofmeyr Stephanie Forrest Patrik D’haeseleer Dept. of Computer Science University.
The Immune System By Mikasa Modi and Anthony Chong.
Specific Cellular Defence.  Range of white blood cells (WBCs) circulate monitoring for damage, pathogens or cancerous cells  In response to damage or.
Specific Defense Mechanisms – The Immune System
Immune System Metaphors Applied to Intrusion Detection and Related Problems by Ian Nunn, SCS, Carleton University
Lecture 14 Immunology: Adaptive Immunity. Principles of Immunity Naturally Acquired Immunity- happens through normal events Artificially Acquired Immunity-
Characteristics of the Specific Immune Response It recognizes and targets “specific” pathogens or foreign substances. It has a “memory,” the capacity to.
Nasraoui, Gonzalez, Cardona, Dasgupta: Scalable Artificial Immune System Based Data Mining NSF-NGDM, Nov. 1-3, 2002, Baltimore, MD Artificial Immune Systems.
Artificial Life “Perhaps the greatest significance of the computer lies in its impact on Man’s view of himself… The computer aids him to obey, for the.
Automatic Test-Data Generation: An Immunological Approach Kostas Liaskos Marc Roper {Konstantinos.Liaskos, TAIC PART 2007.
Zentrum für Hochleistungsrechnen (ZHR) – A Bios Group Presentation Niloy Ganguly Immune System and Search Technology Designing a Fast Search Algorithm.
Specific Immune System
Cell mediated response What the B-cells miss or are not programmed to deal with T-cells are able to handle. Usually 1.
Topics in Artificial Intelligence By Danny Kovach.
APPLICATION OF THE METHOD AND COMBINED ALGORITHM ON THE BASIS OF IMMUNE NETWORK AND NEGATIVE SELECTION FOR IDENTIFICATION OF TURBINE ENGINE SURGING Lytvynenko.
Chapter 24 ~ The Immune System. Animal immune system.
TOLERANCE Deletion, anergy or ignorance?. Normal thyroidDiseased thyroid If the immune system fails to “delete” or anergise tissue reactive lymphocytes,
PP  lymphatic system  spleen  lymphocytes 1. B-cells: wbc that mature in bone marrow 2. T-cells: wbc that mature in thymus.
Specific Defenses of the Host
Human Anatomy and Physiology Immunology: Adaptive defenses.
The Immune System 2 1. Adaptive Specific immunity (acquired) = the ability of the body to defend itself against specific invading agents such as bacteria,
18 Animal Defense Systems Animal defense systems are based on the distinction between self and nonself. There are two general types of defense mechanisms:
Immunology Chapter 43. Innate Immunity Present and waiting for exposure to pathogens Non-specific External barriers and internal cellular and chemical.
Immune System.
+ Immunity: Defense against disease EL: To bring together the learnings on immunity from SAC 4.
Artificial Immune System based Cooperative Strategies for Robot Soccer Competition International Forum on Strategic Technology, p.p , Oct
CONCEPTS OF INFLAMMATION AND THE IMMUNE RESPONSE.
Changing the Rules of the Game Dr. Marco A. Janssen Department of Spatial Economics.
starter CELL comparison Cell Natural Killer Cell B Lymphocyte
T cells Abul K. Abbas: Basic Immunology page (fig3.7, 3.9, 3.11, 3.16 are not required) and (fig 5.11, 5.18 are not required)
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 2005.ACM GECCO.8.Discriminating and visualizing anomalies.
Immunity. Body Defenses First line - barriers Skin and mucous membranes Flushing action –Antimicrobial substances Lysozyme, acids, salts, normal microbiota.
Presentation By SANJOG BHATTA Student ID : July 1’ 2009.
Surface Defect Inspection: an Artificial Immune Approach Dr. Hong Zheng and Dr. Saeid Nahavandi School of Engineering and Technology.
Lec 4 Basic immunology Romana Siddique The Lymphocytes and it’s role in Adaptive immunity.
Chapter 43 The Immune System.
Unit 4 - Immunology and Public Health
Immunology & Public Health
Immune System Chapter 14.
Acquired Immunity.
Immunology & Public Health
Adaptive (specific ) Immunity
Video Immune System Explained- Resources page.
Lec.10 Immune response كلية المأمون الجامعة\قسم تقنيات التحليلات المرضية مادةالمناعة-النظري/المرحلةالثالثة م.م.رشد اياد عبدالحميد.
Immunocomputing and Artificial Immune Systems
Presentation transcript:

1 BY: Nazanin Asadi Zohre Molaei Isfahan University of Technology

2 Outline  History  Natural Immune System  Artificial Immune System  Application  Experiment Result  Reference

History  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  Hunt et al, mid 1990’s – Machine learning 3

4 Basic Immunology

Role of the Immune System  Protect our bodies from infection  Primary immune response Launch a response to invading pathogens  Secondary immune response Remember past encounters Faster response the second time around  The IS is adaptable (presents learning and memory) 5

Where is it ? 6

Lymphocytes  Carry antigen receptors that are specific They are produced in the bone marrow through random re- arrangement  B and T Cells are the main actors of the adaptive immune system 7

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 called the epitope.  Recognition is not just by a single antibody, but a collection of them Learn not through a single agent, but multiple ones 8

T-cells  Regulation of other cells  Active in the immune response Helper T-cells Killer T-cells 9

Immune Responses 10

The Immune System models The are many different viewpoints These views are not mutually exclusive classical networkdanger 11

12 Artificial Immune Systems

Basic concepts  trained detectors(artificial lymphocytes) that detect nonself patterns  need a good repository of self patterns or self and non-self patterns to train ALCs to be self tolerant  need to measure the affinity between an ALC and a pattern  To be able to measure affinity, the representation of the patterns and the ALCs need to have the same structure  The affinity between two ALCs needs to be measured  memory that frequently detect non-self patterns  When an ALC detects non-self patterns, it can be cloned and the clones can be mutated to have more diversity in the search space 13

AIS Framework 14 Algorithms Affinity Representation Application Solution AIS Shape-Space Binary Integer Real-valued Symbolic

Representation – Shape Space  Used for modeling antibody and antigen  Determine a measure to calculate affinity  Hamming shape space(binary)  1 if Ab i != Ag i : 0 otherwise (XOR operator) 15

Representation  Assume the general case:  Ab =  Ab1, Ab2,..., AbL   Ag =  Ag1, Ag2,..., AgL   Binary representation Matching by bits  Continuous (numeric) Real or Integer, typically Euclidian  Symbolic (Categorical /nominal) E.g female or male of the attribute Gender. 16

AIS Framework 17 Algorithms Affinity Representation Application Solution AIS Euclidean Manhattan Hamming

Affinity  Euclidean  Manhattan  Hamming 18

AIS Framework 19 Algorithms Affinity Representation Application Solution AIS Bone Marrow Models Negative Selection Clonal Selection Positive Selection Immune Network Models

Basic AIS Algorithm 20

Negative Selection Algorithms  Forrest 1994: Idea taken from the negative selection of T- cells in the thymus  Applied initially to computer security  Split into two parts: Censoring Monitoring 21

 All patterns and ALCs : as nominal valued attributes or as binary strings  Affinity : r-continuous matching rule  Training set : self patterns 22

Training ALCs with negative selection 23

Clonal Selection 24

Clonal Selection  selection of a set of ALCs with the highest calculated affinity with a non-self pattern  cloned and mutated  compete with the existing set of ALCs  to be exposed to the next non-self pattern Continuous (numeric) 25

ClONALG algorithm  De Castro and Von presented CLONALG as an algorithm,2001  initially proposed to perform machine-learning pattern recognition  Adapted to be applied to optimization problem 26

ClONALG algorithm  main immune aspects taken into account to develop the algorithm maintenance of a specific memory set selection and cloning of the most stimulated Antibodies death of non-stimulated antibodies affinity maturation and re-selection of the clones proportionally to their antigenic affinity generation and maintenance of diversity 27

ClONALG  All patterns in binary strings  Training set : non-self patterns  Affinity : Hamming distance, between ALC and non-self pattern  Lower Hamming distance = stronger affinity  Assumption : One memory ALC for each of the patterns that needs to be recognized in training set 28

ClONALG 29

CLONALG optimization case  an objective function g( ⋅ ) must to be optimized (maximized or minimized)  antibody affinity corresponds to the objective function  each antibody Abi represents an element of the input space  it is no longer necessary to maintain a separate memory set Ab 30

CLONALG optimization case 31

CLONALG optimization case 32

Immune Network Models  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 of an ALC based on the calculated affinity between the ALC and the non-self pattern the calculated affinity between the ALC and network ALCs as co-stimulation and/or co-suppression. 33

Artificial Immune Network  Timmis and Neal,2000  Application  clustering  data visualization  control  optimization domains  AINE defines the new concept of artificial recognition balls (ARBs) population of ARBs links between the ARBs a set of antigen training patterns Some clonal operations for learning 34

Artificial Immune Network 35

Artificial Immune Network  all training patterns in set DT are presented to the set of ARBs  After each iteration, each ARB calculates its stimulation level Allocates resources (i.e. B-Cells) based on its stimulation level as  The stimulation level antigen stimulation network stimulation network suppression 36

Artificial Immune Network 37

Stimulation level  38

Resource allocation 39

Danger Theory Models  distinguishes between what is dangerous and non- dangerous  Include a signal to determine whether a non- self pattern is dangerous or not 40

An Adaptive Mailbox  classifies interesting from uninteresting s  initialization phase (training)  running phase (testing) 41

initialization phase 42

running phase

Application of AIS  network intrusion and anomaly detection  data classification models  virus detection  concept learning  data clustering  robotics  pattern recognition and data mining  optimization of multi-modal functions 44

PSO and AIS 45

PSO and AIS  PSO performs about 56 percent faster than.  AIS performs faster than PSO (about 14 percent) for simpler mathematical functions 46

Reference  Computational Intelligence An introduction, Adndries P.Engelbrecht  Learning and Optimization Using the Clonal Selection Principle, Leandro N. de Castro,,Fernando J. Von Zuben, IEEE,2002  A Comparative Analysis on the Performance of Particle Swarm Optimization and Artificial Immune Systems for Mathematical Test Functions, 1David F.W. Yap, 2S.P. Koh, 2S.K. Tiong,Australian Journal of Basic and Applied Sciences,

48