2004: Engineering Applications of Artificial Immune Systems

767 views
698 views

Published on

Published in: Technology, Education
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
767
On SlideShare
0
From Embeds
0
Number of Embeds
1
Actions
Shares
0
Downloads
62
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

2004: Engineering Applications of Artificial Immune Systems

  1. 1. Engineering ApplicationsEngineering Applications of Artificial Immuneof Artificial Immune SystemsSystems Leandro Nunes de Castro Fernando José Von Zuben lnunes@unisantos.br; vonzuben@dca.fee.unicamp.br Natural Computing Laboratory / Catholic University of Santos Wernher von Braun Center for Advanced Research Laboratory for Bioinformatics and Bio-Inspired Computing / Unicamp Financial Support: CNPq, FAPESP, FAEP
  2. 2. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune2 (Labs Involved in this Research) Catholic University of Santos Wernher von Braun Center for Advanced Research Laboratory of Bioinformatics and Bioinspired Computing Natural Computing Lab
  3. 3. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune3 Topics  Engineering Problems and Their Challenges  Examples of Engineering Problems  Engineering Applications of AIS: Brief Survey from the Literature  Examples of Engineering Applications of AIS from our Research Labs  Discussion
  4. 4. Part I Engineering Problems and Their Challenges
  5. 5. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune5 Engineering Problems  Real-World Problems  An imprecise and incomplete classification:  Pattern Recognition and Classification  Machine Learning  Data Mining  Search and Optimization  Robotics  Control  Industrial Applications
  6. 6. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune6 Engineering Problems  Some Common Features:  Difficulty in modelling  Poorly defined  Dynamic environments  Large number of variables  Missing or noisy variables (attributes)  Highly nonlinear  Difficulty in finding derivatives  Combinatorial solutions (NP-Complete/NP-Hard)
  7. 7. Part II Examples of Engineering Problems
  8. 8. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune8 Examples of Engineering Problems  Non-Linear Control
  9. 9. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune9 Examples of Engineering Problems  Pattern Recognition
  10. 10. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune10 Examples of Engineering Problems  Autonomous Navigation
  11. 11. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune11 Examples of Engineering Problems  Anomaly Detection
  12. 12. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune12 Examples of Engineering Problems  Scheduling
  13. 13. Part III Engineering Applications of AIS: A Brief Survey from the Literature
  14. 14. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune14 AIS: Brief Survey from the Literature  Pattern Recognition and Classification / Machine Learning / Data Mining:  Spectra Recognition  Dasgupta et al., 1999  Surveillance of Infectious Diseases  Tarakanov et al., 2000  Medical Data Analysis  Carter, 2000  Virus Detection and Elimination  Kephart, 1994  Somayaji et al., 1997  Okamoto & Ishida, 1999  Lamont et al., 1999
  15. 15. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune15 AIS: Brief Survey from the Literature  Pattern Recognition and Classification / Machine Learning / Data Mining:  Computer and Network Security  Kephart, 1994  Hedberg, 1996  Kim & Bentley, 1999a,b  Dasgupta, 1999  Gu et al., 2000  Hofmeyr & Forrest, 2000  Skormin et al., 2001  Anchor et al., 2002  Dasgupta & González, 2002  Wang & Hirsbrunner, 2002  de Paula et al., 2004
  16. 16. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune16 AIS: Brief Survey from the Literature  Pattern Recognition and Classification / Machine Learning / Data Mining:  Time Series Data  Dasgupta & Forrest, 1996  Image Processing and Inspection  Aisu & Mizutani, 1996  McCoy & Devarajan, 1997  Sathyanath & Sahin, 2001  Bendiab et al., 2003  Web Mining  Lee et al., 2003  Secker et al., 2003  Oda & White, 2003
  17. 17. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune17 AIS: Brief Survey from the Literature  Pattern Recognition and Classification / Machine Learning / Data Mining:  Fault (Anomaly) Detection  Ishida, 1990  Kayama et al., 1995  Xanthakis et al., 1996  D’haeseleer et al., 1996  Bradley & Tyrrell, 2000  Shulin et al., 2002  Taylor & Corne, 2003  González et al., 2003  Esponda et al., 2003  Kaers et al., 2003  Araujo et al., 2003  Branco et al., 2003
  18. 18. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune18 AIS: Brief Survey from the Literature  Pattern Recognition and Classification / Machine Learning / Data Mining:  Machine Learning  Watkins, 2001  Hunt & Cooke, 1996  Hightower et al., 1996  Potter & de Jong, 1998  Bersini, 1999  Nagano & Yonezawa, 1999  Timmis & Neal, 2001  de Castro & Von Zuben, 2001  Watkins et al., 2004
  19. 19. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune19 AIS: Brief Survey from the Literature  Pattern Recognition and Classification / Machine Learning / Data Mining:  Associative Memory  Gibert & Routen, 1994  Abbattista et al., 1996  Recommender System  Cayzer & Aickelin, 2004  Inductive Problem Solving  Slavov & Nikolaev, 1998  Bankruptcy Prediction  Cheh, 2002
  20. 20. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune20 AIS: Brief Survey from the Literature  Pattern Recognition and Classification / Machine Learning / Data Mining:  Clustering/Classification  Nicosia et al., 2001  Timmis, 2001  de Castro & Timmis, 2002a  Neal, 2002  Zhao & Huang, 2002  Greensmith & Cayzer, 2003  Ceong et al., 2003  Di & Xuefeng, 2003  Nasaroui et al., 2003
  21. 21. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune21 AIS: Brief Survey from the Literature  Pattern Recognition and Classification / Machine Learning / Data Mining:  Bioinformatics  Recognition of promoter sequences: Cooke & Hunt, 1995  Protein structure prediction: Michaud et al., 2001  Spectra classification: Lamont et al., 2004  Gene expression data analysis: Bezerra & de Castro, 2003; Ando & Iba, 2003  Analysis of biological systems: Roy et al., 2002  Bioarrays: Tarakanov et al., 2002
  22. 22. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune22 AIS: Brief Survey from the Literature  Search and Optimization:  Numerical Function Optimization  Mori et al., 1993  Bersini & Varela, 1990  Chun et al., 1998  Huang, 2000  Gaspar & Hirsbrunner, 2002  de Castro & Von Zuben, 2002  de Castro & Timmis, 2002b  Hong & Zong-Yuan, 2002  Walker & Garrett, 2003
  23. 23. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune23 AIS: Brief Survey from the Literature  Search and Optimization:  Constrained Optimization  Hajela & Yoo, 1999  Inventory Optimization  Joshi, 1995  Time Dependent Optimization  Gaspar & Collard, 2000
  24. 24. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune24 AIS: Brief Survey from the Literature  Search and Optimization:  Combinatorial Optimization  Mori et al., 1997, 1998  Endoh et al., 1998  Toma et al., 1999  Hart & Ross, 1999  King et al., 2001  Cui et al., 2001  Costa et al., 2002  Coello Coello et al., 2003  Cutello et al., 2003  Koko et al., 2003
  25. 25. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune25 AIS: Brief Survey from the Literature  Robotics:  Autonomous Navigation  Watanabe et al., 1999  Michelan & Von Zuben, 2002  Hart et al., 2003  Vargas et al., 2003  Canham et al., 2003  Collective Behavior  Mitsumoto et al., 1996  Lee & Sim, 1997  Walking Robots  Ishiguro et al., 1998
  26. 26. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune26 AIS: Brief Survey from the Literature  Control:  Identification, Synthesis and Adaptive Control  Bersini, 1991  Ishida & Adachi, 1996  Krishnakumar & Neidhoefer, 1999  Ding & Ren, 2000  Kim, 2001  Lau & Wong, 2003  Sequential Control  Ootsuki & Sekiguchi, 1999  Feedback Control  Takahashi & Yamada, 1997
  27. 27. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune27  The references from this brief and incomplete survey can be found at:  www.dca.fee.unicamp.br/~lnunes/AIS.html  An extensive and constantly updated bibliography on AIS can be found at:  http://ais.cs.memphis.edu/papers/ais_bibliography.pdf AIS: Brief Survey from the Literature
  28. 28. Part IV Examples of Engineering Applications of AIS from Our Labs
  29. 29. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune29 Applications of AIS from our Labs  Search and Optimization  Multimodal search  Dynamic environments*  Blind equalization  Pattern Recognition and Classification  Classification and clustering  Detection of buffer overflow attacks
  30. 30. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune30 Applications of AIS from our Labs  Machine Learning  Neural network initialization  Neural network training*  Hybrid neural networks  Robotics  Autonomous navigation  Bioinformatics  Gene expression data analysis
  31. 31. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune31 Search and Optimization  Multimodal Search  CLONALG (de Castro & Von Zuben, 2002) Pr M Select Clone Pn C C* (1) (2) (3) (5) Re-select Nd (6) Maturate (4)
  32. 32. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune32 Search and Optimization  Multimodal Search  CLONALG (de Castro & Von Zuben, 2002) CLONALG Standard GA
  33. 33. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune33 Search and Optimization  Combinatorial Search  CLONALG (de Castro & Von Zuben, 2002) 7 1 8 14 2 15 3 4 11 12 13 17 23 27 30 26 19 21 24 29 28 25 22 20 18 16 6 9 10 5 TSP - 300 gen
  34. 34. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune34 Search and Optimization  Combinatorial Search  Copt-aiNet (de Sousa et al., 2004)
  35. 35. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune35 Search and Optimization  CLONALG (de Castro & Von Zuben, 2002) DEMO 1: CLONALGDEMO 1: CLONALG
  36. 36. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune36 Search and Optimization  Multimodal Search  opt-aiNet (de Castro & Timmis, 2002b)  The algorithm for opt-aiNet is an adaptation of a discrete artificial immune network usually applied in data analysis  Features of opt-aiNet:  population size dynamically adjustable  exploitation and exploration of the search-space  capability of locating multiple optima  automatic stopping criterion
  37. 37. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune37 Search and Optimization  Multimodal Search  opt-aiNet (de Castro & Timmis, 2002b) 1. Initialize population (initial number not relevant) 2. While not [constant memory population], do 2.1 Calculate fitness and generate clones for each network cell. 2.2 Mutate clones proportionally to fitness and determine the fitness again. 2.3 Calculate the average fitness. 2.4 If average fitness does not vary, then continue. Else, return to step 2.1 2.5 Calculate the affinity among cells and suppress all but one whose affinities are less than the suppression threshold σs and determine the number of network cells after suppression. 2.6 Introduce a percentage of randomly generated cells and return to step 2. 3. EndWhile
  38. 38. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune38 Search and Optimization  Multimodal Search  opt-aiNet (de Castro & Timmis, 2002b)
  39. 39. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune39 Search and Optimization  Communications Engineering  Search for the optimal Wiener equalizer  opt-aiNet (Attux et al., 2003)  Optimal Wiener Equalizer  y(n) = wT .x(n)  JW = E{[s(n-d) – y(n)]2 } CHANNEL EQUALIZER s(n) x(n) y(n)
  40. 40. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune40 Search and Optimization  Optimal Wiener Equalizer  The Constant Modulus (CM) criterion is used for blind equalization  To find the CM global optimum is equivalent to determining the optimal Wiener solution (best equalizer)  CM results in a multi-modal problem  JCM = E{[R2 - |y(n)|2 ]2 } [ ] [ ]2 4 2 )( )( nsE nsE R =
  41. 41. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune41 Search and Optimization  Optimal Wiener Equalizer via CM Search  Sample performance  HC1 = 1 + 0.4z-1 + 0.9z-2 + 1.4z-3
  42. 42. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune42 Search and Optimization  opt-aiNet (de Castro & Timmis, 2002b) DEMO 2: opt-aiNetDEMO 2: opt-aiNet
  43. 43. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune43 Pattern Recognition  Classification and Clustering  CLONALG (de Castro & Von Zuben, 2002) (a) Input patterns (b) 0 generations (c) 50 generations (d) 100 generations (e) 200 generations
  44. 44. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune44 Pattern Recognition  Classification and Clustering  aiNet (de Castro & Von Zuben, 2001)  Definition:  aiNet is an edge-weighted graph, not necessarily fully connected, composed of a set of nodes and sets of node pairs with a weight assigned specified to each connected edge.  Features:  knowledge distributed among cells  competitive learning (unsupervised)  constructive model with pruning phases  generation and maintenance of diversity
  45. 45. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune45 Pattern Recognition  aiNet:  Growing:  clonal selection principle  Learning:  directed affinity maturation  Pruning:  immune network theory
  46. 46. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune46 Pattern Recognition  aiNet at each generation:  For each Ag  Affinity with the antigen (Ai) Agi-Ab  Clonal selection (n cells) ∝ Ai  Cloning ∝ Ai  Directed maturation (mutation) ∝ 1/Ai  Re-selection (ζ%) ∝ Ai  Natural death (σd) ∝ 1/Ai  Affinity between the network cells (Dii) Ab-Ab  Clonal suppression (σs) ∝ Dii : (m - memory)  Mt ← [Mt;m]  Network suppression (σs) ∝ Dii : (M ⊆ Mt)  M ← [M;meta]
  47. 47. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune47 Pattern Recognition  Clustering 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 x y Training Patterns 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Final Network Structure
  48. 48. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune48 Pattern Recognition  Clustering -2 -1 0 1 2 -2 0 2 4 -1.5 -1 -0.5 0 0.5 1 1.5 -1 -0.5 0 0.5 1 -1 0 1 2 3 -1 -0.5 0 0.5 1 1.5 Final Network Structure
  49. 49. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune49 Pattern Recognition  The Immune Response of aiNet (de Castro, 2004)  Network Hypotheses Used in aiNet  Clonal selection, expansion and maturation to foreign stimulation  Network interactions (suppression)  Network metadynamics  Dynamics: Mix between learning and evolution Abk* = Abk + αk (Ag – Abk); αk ∝ Affk; k = 1,...,Nc. ∆Ab = Nc − Ns + Nb − Nd
  50. 50. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune50 Pattern Recognition  The Immune Response of aiNet
  51. 51. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune51 Pattern Recognition  The Immune Response of aiNet 0 200 400 600 800 1000 1200 1400 1600 1800 0.86 0.88 0.9 0.92 0.94 0.96 0.98 1
  52. 52. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune52 Pattern Recognition  The Immune Response of aiNet 0 50 100 150 200 250 10 1 10 2 10 3 10 4 Primary, Secondary and Cross-Reactive Immune Responses Iteration AntibodyConcentration Ag1 Ag1, Ag11, Ag2 Responses to Ag1 Response to Ag2 Response to Ag11
  53. 53. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune53 Pattern Recognition  aiNet (de Castro & Von Zuben, 2001) DEMO 3: aiNetDEMO 3: aiNet
  54. 54. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune54 Pattern Recognition  Detection of Buffer Overflow Attacks  ADENOIDS (de Paula et al., 2004)  An ID framework inspired by the architecture of the immune system  Prototype of an IDS based on the proposed framework  Elaborates on some ideas from Aickelin et al., 2002 about the Danger Theory as a missing link for AIS
  55. 55. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune55 Pattern Recognition  Danger Theory (Matzinger, 2002)
  56. 56. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune56 Pattern Recognition  Detection of Buffer Overflow Attacks  Desirable features based on the immune system (danger theory)  Automated intrusion recovery  Attack signature extraction  Potential to improve behavior-based detection
  57. 57. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune57 Pattern Recognition  Framework for Intrusion Detection
  58. 58. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune58 Machine Learning  Neural Network Initialization  SAND (de Castro & Von Zuben, 2001a)  Initial neural network (NN) weights:  learning speed  generalization performance  Correlation: initial set of weights and initial repertoire of immune cells and molecules  SAND: a Simulated ANnealing model to increase population Diversity
  59. 59. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune59 Machine Learning  Neural Network Initialization  Affinity measure:  Proposed cost (energy) function: ∑ = −= L i ii yxED 1 2 )( ∑ = = N i i N 1 1 II ( ) 2/1 IIT R = )1(100(%) RE −×=  Average unit vector  Resultant vector (distance from the origin of the coordinate system)  Percentage energy
  60. 60. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune60 Machine Learning  Neural Network Initialization  Ab → weight vector  Diverse antibodies in ℜL → neurons with well distributed weight vectors in ℜL  SAND is applied separately to each network layer  The vectors (Ab) have unitary norms and can be normalized to avoid neuron saturation
  61. 61. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune61 Machine Learning  Neural Network Initialization OLS: Striped INIT: Red SAND: Blank
  62. 62. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune62 Machine Learning  RBF Neural Network Center Selection  de Castro & Von Zuben, 2001  The performance of the RBF neural network depends on the number, positions and dispersions of the basis functions composing the network hidden layer  Traditional methods:  randomly choose input vectors from the training data set;  vectors obtained from unsupervised clustering algorithms;  vectors obtained by supervised learning schemes.
  63. 63. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune63 Machine Learning  RBF Neural Network Center Selection  Solution based on aiNet
  64. 64. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune64 Machine Learning  RBF Neural Network Center Selection (1) (2) (3) -1.5 -1 -0.5 0 0.5 1 -π/2 π/2 1.5
  65. 65. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune65 Machine Learning  RBF Neural Network Center Selection -0.5 0 0.5 1 1.5 2 2.5 -0.5 0 0.5 1 1.5 2 2.5 ICS Iris data set Best performance
  66. 66. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune66 Machine Learning  Boolean Neural Network (ABNET)  de Castro et al., 2003  Main Features:  clustering, or grouping of similar patterns  capability of solving binary tasks  growing learning with pruning phases  Main loop of the algorithm  Choose randomly an antigen (pattern)  Determine the cell Abk with highest affinity  Update the weight vector of this cell  Increase the concentration level (τj) of this cell  Attribute va = k
  67. 67. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune67 Machine Learning  ABNET Ab population Ab re-selection Clone Death Affinity maturation Most stimulated cell Non-stimulated cell Ab selection Antigenic stimuli Neurons (k) Competition Split Prune Weight update Winner Inactive neuron Input patterns
  68. 68. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune68 Machine Learning  ABNET  Binary character recognition 0 1 2 3 4 5 0.4 0.3 0.2 0.1 0 20 40 60 80 100 ε NoiseTolerance-ABNET Noiselevel Classification(%) 2) Cross-reactivity (generalization) (a) ε ≥ 13.75% Noise tolerance: (b) ε ≥ 13.75%
  69. 69. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune69 Machine Learning  ABNET  Animals data set ABNET(0-valuedweightsomitted) Lion Tiger Wolf Dog Fox Cat Horse/Zebra Cow Owl/Hawk Dove Hen Duck Goose Eagle Mammals Birds
  70. 70. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune70 Machine Learning  ABNET (de Castro et al., 2003) DEMO 4: ABNETDEMO 4: ABNET
  71. 71. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune71 Robotics  Autonomous Navigation based on AIS  Michelan & Von Zuben, 2002  Based on the works:  Ishiguro et al., 1997; Farmer et al., 1986
  72. 72. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune72 Robotics  Autonomous Navigation based on AIS  Autonomous control system of a mobile robot based on the immune network theory  Each network node corresponds to a specific antibody and describes a particular control action for the robot  The antigens are the current state of the robot  The network dynamics corresponds to the variation of antibody concentration levels, which change according to both mutual interaction of antibody nodes and of antibodies and antigens  It is proposed an evolutionary mechanism to determine the network configuration
  73. 73. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune73 Robotics  Autonomous Navigation based on AIS
  74. 74. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune74 Robotics  Autonomous Navigation based on AIS  Objectives of navigation  Antibody structure garbageEcollisionEstepEtEtE −−−−= )1()(
  75. 75. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune75 Robotics  Autonomous Navigation based on AIS  Network example
  76. 76. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune76 Robotics  Autonomous Navigation based on AIS  Dynamics )()()( )( 11 takmtamtam dt tdA i N k iikik N j jji i         −+−= ∑∑ == ))(5.0exp(1 1 )1( tA ta i i −+ =+
  77. 77. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune77 Robotics  Autonomous Navigation based on AIS Immune network Evolved network
  78. 78. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune78 Robotics  Autonomous Navigation based on AIS  Implementation on Khepera II® : Vargas et al., 2003
  79. 79. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune79 Robotics  Autonomous Navigation (Vargas et al., 2003) DEMO 5: Collision AvoidanceDEMO 5: Collision Avoidance
  80. 80. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune80 Bioinformatics  Gene Expression Data Analysis  Bezerra & de Castro, 2003  de Sousa et al., 2004  The Problem  Clustering gene expression data  Recent approach in bioinformatics that surged with the development of the DNA Microarrays  DNA Microarrays  Experimental technique that measures the expression level of many genes simultaneously  A quantitative change in the scale of the experiments led to a qualitative change in the analyses, where the genes may be studied under a genome wide perspective
  81. 81. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune81 Bioinformatics  Gene Expression Data Analysis  Genes belonging to the same cluster may, among other things  Share the same regulatory system  Have similar properties or functions  Code products that interact physically  Experimental Data  Gene expression data of the budding yeast Saccharomyces cerevisiae, obtained from Eisen et al. (1998)  Total of 2467 genes in 79 different conditions
  82. 82. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune82 Bioinformatics  Gene Expression Data Analysis  Clusters initially analyzed C, E, F and H (68 genes)  Results with full set: No natural cluster
  83. 83. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune83 Bioinformatics  Multiple Simultaneous Views
  84. 84. Part V Discussion
  85. 85. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune85 Discussion  Vast number of applications available  Great potential for further applications and developments  Some issues that still deserve investigation:  Formal aspects  Comparison (theoretical and empirical) with other approaches  Loads of testing  Real benefits (Are they really useful?)  Danger theory  How far to stretch the metaphor?
  86. 86. ICARIS 2004 - Engineering Applications of AIS - Leandro Nune86 Discussion  Current trends in our labs  Improvements on the many versions of aiNet  Optimization on dynamic environments  Bioinformatics, mainly gene expression data analysis  Feedforward neural network training  Danger theory  Anomaly detection  This Tutorial on the Web:  www.dca.fee.unicamp.br/~lnunes/AIS.html

×