Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.
A Matlab Tour on Some AIS Algorithms BIC 2005:  International Symposium on Bio-Inspired Computing Johor, MY, 10 th  Septem...
<ul><li>CLONALG: A Clonal Selection Algorithm </li></ul><ul><li>aiNet: An Artificial Immune Network </li></ul><ul><li>ABNE...
CLONALG A Clonal Selection Algorithm
<ul><li>Increasing interest in biologically inspired systems </li></ul><ul><li>Systemic view of the immune system </li></u...
Clonal Selection Principle
Continuous Learning
Affinity Maturation <ul><li>The cells that are most stimulated by the antigens suffer a  hypermutation  process </li></ul>...
Hypermutation   Editing
CLONALG: Block Diagram
<ul><ul><ul><li>1) Generate a set (P) of candidate solutions, composed of the subset of memory cells (M) added to the rema...
Test Problem I <ul><li>Pattern recognition (learning) </li></ul><ul><li>Cross-reactivity (generalization capability) </li>...
<ul><li>Pattern Recognition (Learning) </li></ul>CLONALG - Performance I 00 generations 10 generations 20 generations 50 g...
<ul><li>Optimization </li></ul><ul><ul><li>function maximization </li></ul></ul>Test Problem II 200 individuals  randomly ...
<ul><li>Multimodal Optimization (Maximization) </li></ul><ul><ul><li>Comparison with the Standard Genetic Algorithm (GA) <...
<ul><li>Optimization </li></ul><ul><ul><li>Travelling Salesman Problem (TSP) </li></ul></ul>Test Problem III Cities     Op...
<ul><li>Travelling Salesman (TSP) (Minimization) </li></ul>CLONALG - Performance III
CLONALG: Discussion <ul><li>General purpose algorithm inspired by the clonal selection and affinity maturation processes <...
aiNet An Artificial Immune Network
Immune Network Theory
aiNet: Basic Principles (I) <ul><li>Definition:  </li></ul><ul><ul><li>The evolutionary artificial immune network, named a...
<ul><li>Features: </li></ul><ul><ul><li>knowledge distributed among the cells </li></ul></ul><ul><ul><li>competitive learn...
aiNet: Training Algorithm <ul><li>At each generation: </li></ul><ul><ul><li>For each Ag </li></ul></ul><ul><ul><ul><li>  ...
<ul><ul><ul><li> (affinity) </li></ul></ul></ul><ul><ul><ul><li>     (clonal selection) </li></ul></ul></ul><ul><ul><ul>...
Test Problem I <ul><li>Five Linearly Separable Classes </li></ul>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...
aiNet - Performance I Minimal Spanning Tree 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 Number of Clusters (Valleys) 1 1 2 ...
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 aiNet - Performance I
<ul><li>2-Donuts: 500 samples </li></ul>Test Problem II -2 -1 0 1 2 -2 0 2 4 -1.5 -1 -0.5 0 0.5 1 1.5
Number of Clusters (Valleys) Minimal Spanning Tree aiNet - Performance II 1 0 5 10 15 20 25 30 35 0 0.1 0.2 0.3 0.4 0.5 0....
-1 -0.5 0 0.5 1 -1 0 1 2 3 -1 -0.5 0 0.5 1 1.5 Final Network Structure aiNet - Performance II
aiNet: Discussion <ul><li>Iterative learning </li></ul><ul><li>Robustness with low redundancy (data compression) </li></ul...
ABNET An Antibody Network
ABNET <ul><li>A single-layer feedforward neural network trained using ideas from the immune system </li></ul><ul><li>Const...
ABNET: Basic Functioning (I)
<ul><li>Affinity measure (Hamming distance): </li></ul><ul><li>Main loop of the algorithm </li></ul><ul><ul><li>Choose ran...
ABNET: Growing
ABNET: Pruning
ABNET: Weight Update
ABNET - Performance 2) Cross-reactivity (generalization) (a)       13.75% Noise tolerance: (b)       13.75%
ABNET: Discussion <ul><li>Performs clustering (data reduction) </li></ul><ul><li>Easily implemented in hardware </li></ul>...
Opt-aiNet An Optimization version of aiNet
Introduction <ul><li>The algorithm for opt-aiNet is an adaptation of a discrete artificial immune network usually applied ...
Immune Networks <ul><li>N. Jerne suggested that immune cells and molecules present antigenic peptides </li></ul>
opt-aiNet <ul><li>1.  Randomly initialize a population of cells (initial number not relevant) </li></ul><ul><li>2.  While ...
Related Strategies <ul><li>CLONALG: </li></ul><ul><ul><li>encoding, static population size, no inter-cell interaction, dif...
Simulation Results (I) <ul><li>Multi Function </li></ul>
Simulation Results (II) <ul><li>Roots Function </li></ul>
Simulation Results (III) <ul><li>Schaffer’s Function </li></ul>
Opt-aiNet: Discussion <ul><li>The algorithm is an adaptation of an immune network model designed to perform data analysis ...
Final Comments <ul><li>Biological Inspiration </li></ul><ul><ul><li>utility and extension </li></ul></ul><ul><ul><li>impro...
<ul><li>CLONALG </li></ul><ul><ul><li>high degree of parallelism </li></ul></ul><ul><ul><li>by controlling the hipermutati...
<ul><li>aiNet </li></ul><ul><ul><li>Models continuous spaces without the need of integration </li></ul></ul><ul><ul><li>It...
Final Comments <ul><li>Opt-aiNet </li></ul><ul><ul><li>;lsdkasdkj </li></ul></ul>
[email_address] Questions?  Comments?
Upcoming SlideShare
Loading in …5
×

2005: A Matlab Tour on Artificial Immune Systems

6,783 views

Published on

BIC 2005 (Biologically Inspired Computing Conference), Johor, Malaysia

Published in: Technology, Education
  • DOWNLOAD THIS BOOKS INTO AVAILABLE FORMAT (Unlimited) ......................................................................................................................... ......................................................................................................................... Download Full PDF EBOOK here { https://tinyurl.com/yyxo9sk7 } ......................................................................................................................... Download Full EPUB Ebook here { https://tinyurl.com/yyxo9sk7 } ......................................................................................................................... ACCESS WEBSITE for All Ebooks ......................................................................................................................... Download Full PDF EBOOK here { https://tinyurl.com/yyxo9sk7 } ......................................................................................................................... Download EPUB Ebook here { https://tinyurl.com/yyxo9sk7 } ......................................................................................................................... Download doc Ebook here { https://tinyurl.com/yyxo9sk7 } ......................................................................................................................... ......................................................................................................................... ......................................................................................................................... .............. Browse by Genre Available eBooks ......................................................................................................................... Art, Biography, Business, Chick Lit, Children's, Christian, Classics, Comics, Contemporary, Cookbooks, Crime, Ebooks, Fantasy, Fiction, Graphic Novels, Historical Fiction, History, Horror, Humor And Comedy, Manga, Memoir, Music, Mystery, Non Fiction, Paranormal, Philosophy, Poetry, Psychology, Religion, Romance, Science, Science Fiction, Self Help, Suspense, Spirituality, Sports, Thriller, Travel, Young Adult,
       Reply 
    Are you sure you want to  Yes  No
    Your message goes here
  • DOWNLOAD FULL. BOOKS INTO AVAILABLE FORMAT, ......................................................................................................................... ......................................................................................................................... 1.DOWNLOAD FULL. PDF EBOOK here { https://tinyurl.com/y8nn3gmc } ......................................................................................................................... 1.DOWNLOAD FULL. EPUB Ebook here { https://tinyurl.com/y8nn3gmc } ......................................................................................................................... 1.DOWNLOAD FULL. doc Ebook here { https://tinyurl.com/y8nn3gmc } ......................................................................................................................... 1.DOWNLOAD FULL. PDF EBOOK here { https://tinyurl.com/y8nn3gmc } ......................................................................................................................... 1.DOWNLOAD FULL. EPUB Ebook here { https://tinyurl.com/y8nn3gmc } ......................................................................................................................... 1.DOWNLOAD FULL. doc Ebook here { https://tinyurl.com/y8nn3gmc } ......................................................................................................................... ......................................................................................................................... ......................................................................................................................... .............. Browse by Genre Available eBooks ......................................................................................................................... Art, Biography, Business, Chick Lit, Children's, Christian, Classics, Comics, Contemporary, Cookbooks, Crime, Ebooks, Fantasy, Fiction, Graphic Novels, Historical Fiction, History, Horror, Humor And Comedy, Manga, Memoir, Music, Mystery, Non Fiction, Paranormal, Philosophy, Poetry, Psychology, Religion, Romance, Science, Science Fiction, Self Help, Suspense, Spirituality, Sports, Thriller, Travel, Young Adult,
       Reply 
    Are you sure you want to  Yes  No
    Your message goes here
  • DOWNLOAD FULL BOOKS INTO AVAILABLE FORMAT ......................................................................................................................... ......................................................................................................................... 1.DOWNLOAD FULL PDF EBOOK here { https://tinyurl.com/y8nn3gmc } ......................................................................................................................... 1.DOWNLOAD FULL EPUB Ebook here { https://tinyurl.com/y8nn3gmc } ......................................................................................................................... 1.DOWNLOAD FULL doc Ebook here { https://tinyurl.com/y8nn3gmc } ......................................................................................................................... 1.DOWNLOAD FULL PDF EBOOK here { https://tinyurl.com/y8nn3gmc } ......................................................................................................................... 1.DOWNLOAD FULL EPUB Ebook here { https://tinyurl.com/y8nn3gmc } ......................................................................................................................... 1.DOWNLOAD FULL doc Ebook here { https://tinyurl.com/y8nn3gmc } ......................................................................................................................... ......................................................................................................................... ......................................................................................................................... .............. Browse by Genre Available eBooks ......................................................................................................................... Art, Biography, Business, Chick Lit, Children's, Christian, Classics, Comics, Contemporary, Cookbooks, Crime, Ebooks, Fantasy, Fiction, Graphic Novels, Historical Fiction, History, Horror, Humor And Comedy, Manga, Memoir, Music, Mystery, Non Fiction, Paranormal, Philosophy, Poetry, Psychology, Religion, Romance, Science, Science Fiction, Self Help, Suspense, Spirituality, Sports, Thriller, Travel, Young Adult,
       Reply 
    Are you sure you want to  Yes  No
    Your message goes here

2005: A Matlab Tour on Artificial Immune Systems

  1. 1. A Matlab Tour on Some AIS Algorithms BIC 2005: International Symposium on Bio-Inspired Computing Johor, MY, 10 th September 2005 Dr. Leandro Nunes de Castro [email_address] http://lsin.unisantos.b/lnunes Catholic University of Santos - UniSantos/Brazil
  2. 2. <ul><li>CLONALG: A Clonal Selection Algorithm </li></ul><ul><li>aiNet: An Artificial Immune Network </li></ul><ul><li>ABNET: An Antibody Network </li></ul><ul><li>Opt-aiNet: An Optimization Version of aiNet </li></ul><ul><li>Discussion </li></ul>Outline
  3. 3. CLONALG A Clonal Selection Algorithm
  4. 4. <ul><li>Increasing interest in biologically inspired systems </li></ul><ul><li>Systemic view of the immune system </li></ul><ul><li>Main goals: </li></ul><ul><ul><li>Provide a better understanding of the immune system </li></ul></ul><ul><ul><li>Solve engineering problems </li></ul></ul><ul><ul><li>Study immune learning and memory </li></ul></ul>CLONALG
  5. 5. Clonal Selection Principle
  6. 6. Continuous Learning
  7. 7. Affinity Maturation <ul><li>The cells that are most stimulated by the antigens suffer a hypermutation process </li></ul><ul><ul><li>single point, short deletions and sequence exchange </li></ul></ul><ul><li>The hypermutation is proportional to antigenic affinity </li></ul><ul><li>The higher the cell affinity with the antigen, the greater its probability of being selected for differentiation and memory, thus surviving longer </li></ul><ul><li>The mutation rate is proportional to antigenic affinity </li></ul><ul><li>The editing process promotes a better exploration of the possible antigenic receptors </li></ul>
  8. 8. Hypermutation  Editing
  9. 9. CLONALG: Block Diagram
  10. 10. <ul><ul><ul><li>1) Generate a set (P) of candidate solutions, composed of the subset of memory cells (M) added to the remaining (P r ) population (P = P r  + M) </li></ul></ul></ul><ul><ul><ul><li>2) Determine the n best individuals of the population (P n ), based on an affinity measure </li></ul></ul></ul><ul><ul><ul><li>3) Clone (reproduce) these n best individuals of the population, giving rise to a temporary population of clones ( C ). The clone size is an increasing function of the affinity with the antigen; </li></ul></ul></ul><ul><ul><ul><li>4) Submit the population of clones to a hypermutation scheme, where the hypermutation is proportional to the affinity of the antibody with the antigen. A maturated antibody population is generated ( C* ); </li></ul></ul></ul><ul><ul><ul><li>5) Re-select the improved individuals from C* to compose the memory set. Some members of the P set can be replaced by other improved members of C* ; </li></ul></ul></ul><ul><ul><ul><li>6) Replace d low affinity antibodies of the population, maintaining its diversity. </li></ul></ul></ul>CLONALG: Algorithm
  11. 11. Test Problem I <ul><li>Pattern recognition (learning) </li></ul><ul><li>Cross-reactivity (generalization capability) </li></ul>
  12. 12. <ul><li>Pattern Recognition (Learning) </li></ul>CLONALG - Performance I 00 generations 10 generations 20 generations 50 generations 75 generations 100 generations 150 generations 200 generations 250 generations
  13. 13. <ul><li>Optimization </li></ul><ul><ul><li>function maximization </li></ul></ul>Test Problem II 200 individuals randomly distributed
  14. 14. <ul><li>Multimodal Optimization (Maximization) </li></ul><ul><ul><li>Comparison with the Standard Genetic Algorithm (GA) </li></ul></ul>CLONALG - Performance II GA CLONALG
  15. 15. <ul><li>Optimization </li></ul><ul><ul><li>Travelling Salesman Problem (TSP) </li></ul></ul>Test Problem III Cities Optimal Path (48872 um) 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 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
  16. 16. <ul><li>Travelling Salesman (TSP) (Minimization) </li></ul>CLONALG - Performance III
  17. 17. CLONALG: Discussion <ul><li>General purpose algorithm inspired by the clonal selection and affinity maturation processes </li></ul><ul><li>Capabilities: </li></ul><ul><ul><li>learning and maintenance of high quality memory </li></ul></ul><ul><ul><li>optimization </li></ul></ul><ul><li>Crude version </li></ul><ul><li>GA  CSA: </li></ul><ul><ul><li>same coding schemes </li></ul></ul><ul><ul><li>different sources of inspiration </li></ul></ul><ul><ul><li>related sequence of steps </li></ul></ul>
  18. 18. aiNet An Artificial Immune Network
  19. 19. Immune Network Theory
  20. 20. aiNet: Basic Principles (I) <ul><li>Definition: </li></ul><ul><ul><li>The evolutionary artificial immune network, named aiNet, is an edge-weighted graph , not necessarily fully connected, composed of a set of nodes, called cells , and sets of node pairs called edges with a number assigned called weight , or connection strength , specified to each connected edge. </li></ul></ul>
  21. 21. <ul><li>Features: </li></ul><ul><ul><li>knowledge distributed among the cells </li></ul></ul><ul><ul><li>competitive learning (unsupervised) </li></ul></ul><ul><ul><li>constructive model with pruning phases </li></ul></ul><ul><ul><li>generation and maintenance of diversity </li></ul></ul><ul><li>Growing: </li></ul><ul><ul><li>clonal selection principle </li></ul></ul><ul><li>Learning: </li></ul><ul><ul><li>directed affinity maturation </li></ul></ul><ul><li>Pruning: </li></ul><ul><ul><li>natural death rate (low stimulated cells) </li></ul></ul>aiNet: Basic Principles (II)
  22. 22. aiNet: Training Algorithm <ul><li>At each generation: </li></ul><ul><ul><li>For each Ag </li></ul></ul><ul><ul><ul><li> Affinity with the antigen ( A i ) Ag i -Ab </li></ul></ul></ul><ul><ul><ul><li> Clonal selection ( n cells)  A i </li></ul></ul></ul><ul><ul><ul><li> Cloning  A i </li></ul></ul></ul><ul><ul><ul><li> Directed maturation (mutation)  1/ A i </li></ul></ul></ul><ul><ul><ul><li> Re-selection (  %)  A i </li></ul></ul></ul><ul><ul><ul><li> Natural death (  d )  1/ A i </li></ul></ul></ul><ul><ul><ul><li> Affinity between the network cells ( D ii ) Ab-Ab </li></ul></ul></ul><ul><ul><ul><li> Clonal suppression (  s )  D ii : ( m - memory) </li></ul></ul></ul><ul><ul><ul><li> M t  [M t ; m ] </li></ul></ul></ul><ul><ul><li>Network suppression (  s )  D ii : (M  M t ) </li></ul></ul><ul><ul><li>M  [M;meta] </li></ul></ul>
  23. 23. <ul><ul><ul><li> (affinity) </li></ul></ul></ul><ul><ul><ul><li> (clonal selection) </li></ul></ul></ul><ul><ul><ul><li> (directed mutation) </li></ul></ul></ul><ul><ul><ul><li> (Re-selection) </li></ul></ul></ul><ul><ul><ul><li> (self discrimination) </li></ul></ul></ul><ul><ul><ul><li> (clonal suppression) </li></ul></ul></ul><ul><ul><ul><li>Stopping criterion : or fixed number of generations </li></ul></ul></ul>aiNet: Arithmetic
  24. 24. Test Problem I <ul><li>Five Linearly Separable Classes </li></ul>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
  25. 25. aiNet - Performance I Minimal Spanning Tree 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 Number of Clusters (Valleys) 1 1 2 3 4 5 6 7 8 9 10 11 12 13 14
  26. 26. 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 aiNet - Performance I
  27. 27. <ul><li>2-Donuts: 500 samples </li></ul>Test Problem II -2 -1 0 1 2 -2 0 2 4 -1.5 -1 -0.5 0 0.5 1 1.5
  28. 28. Number of Clusters (Valleys) Minimal Spanning Tree aiNet - Performance II 1 0 5 10 15 20 25 30 35 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
  29. 29. -1 -0.5 0 0.5 1 -1 0 1 2 3 -1 -0.5 0 0.5 1 1.5 Final Network Structure aiNet - Performance II
  30. 30. aiNet: Discussion <ul><li>Iterative learning </li></ul><ul><li>Robustness with low redundancy (data compression) </li></ul><ul><li>Clustering </li></ul><ul><li>Related with neural networks </li></ul><ul><li>User-defined parameters </li></ul><ul><li>Gave rise to a number of other algorithms </li></ul>
  31. 31. ABNET An Antibody Network
  32. 32. ABNET <ul><li>A single-layer feedforward neural network trained using ideas from the immune system </li></ul><ul><li>Constructive architecture with pruning phases </li></ul><ul><li>Boolean weights </li></ul>
  33. 33. ABNET: Basic Functioning (I)
  34. 34. <ul><li>Affinity measure (Hamming distance): </li></ul><ul><li>Main loop of the algorithm </li></ul><ul><ul><li>Choose randomly an antigen (pattern) </li></ul></ul><ul><ul><li>Determine the cell Ab k with highest affinity </li></ul></ul><ul><ul><li>Update the weight vector of this cell </li></ul></ul><ul><ul><li>Increase the concentration level (  j ) of this cell </li></ul></ul><ul><ul><li>Attribute v a = k </li></ul></ul>ABNET: Basic Functioning (I)
  35. 35. ABNET: Growing
  36. 36. ABNET: Pruning
  37. 37. ABNET: Weight Update
  38. 38. ABNET - Performance 2) Cross-reactivity (generalization) (a)   13.75% Noise tolerance: (b)   13.75%
  39. 39. ABNET: Discussion <ul><li>Performs clustering (data reduction) </li></ul><ul><li>Easily implemented in hardware </li></ul><ul><li>Robust to solve binary tasks </li></ul><ul><li>Adapted to solve real-valued problems, both clustering and classification </li></ul>
  40. 40. Opt-aiNet An Optimization version of aiNet
  41. 41. Introduction <ul><li>The algorithm for opt-aiNet is an adaptation of a discrete artificial immune network usually applied in data analysis </li></ul><ul><li>Features of opt-aiNet: </li></ul><ul><ul><li>population size dynamically adjustable </li></ul></ul><ul><ul><li>exploitation and exploration of the search-space </li></ul></ul><ul><ul><li>capability of locating multiple optima </li></ul></ul><ul><ul><li>automatic stopping criterion </li></ul></ul>
  42. 42. Immune Networks <ul><li>N. Jerne suggested that immune cells and molecules present antigenic peptides </li></ul>
  43. 43. opt-aiNet <ul><li>1. Randomly initialize a population of cells (initial number not relevant) </li></ul><ul><li>2. While not [constant memory population], do </li></ul><ul><li>2.1 Calculate the fitness and normalize the vector of fitnesses. </li></ul><ul><li>2.2 Generate a number Nc of clones for each network cell. </li></ul><ul><li>2.3 Mutate each clone proportionally to the fitness of its parent cell, but keep the parent cell. </li></ul><ul><li>2.4 Determine the fitness of all individuals of the population. </li></ul><ul><li>2.5 For each clone, select the cell with highest fitness and calculate the average fitness of the selected population. </li></ul><ul><li>2.6 If the average fitness of the population is not significantly different from the previous iteration, then continue. Else , return to step 2.1 </li></ul><ul><li>2.7 Determine the affinity of all cells in the network. Suppress all but the highest fitness of those cells whose affinities are less than the suppression threshold  s and determine the number of network cells, named memory cells, after suppression. </li></ul><ul><li>2.8 Introduce a percentage d % of randomly generated cells and return to step 2. </li></ul><ul><li>3. EndWhile </li></ul>
  44. 44. Related Strategies <ul><li>CLONALG: </li></ul><ul><ul><li>encoding, static population size, no inter-cell interaction, different mutation scheme </li></ul></ul><ul><li>Evolution Strategies </li></ul><ul><ul><li>equal to (  +  )-ES, where  = N and  = Nc; both use Gaussian mutation, but with different standard deviations, static population size, no diversity introduction, no direct interaction within the population </li></ul></ul>
  45. 45. Simulation Results (I) <ul><li>Multi Function </li></ul>
  46. 46. Simulation Results (II) <ul><li>Roots Function </li></ul>
  47. 47. Simulation Results (III) <ul><li>Schaffer’s Function </li></ul>
  48. 48. Opt-aiNet: Discussion <ul><li>The algorithm is an adaptation of an immune network model designed to perform data analysis </li></ul><ul><li>Features: </li></ul><ul><ul><li>Exploration and exploitation of the search-space </li></ul></ul><ul><ul><li>Double-plastic search </li></ul></ul><ul><ul><li>Automatic convergence criteria </li></ul></ul><ul><li>Adapted to solve combinatorial and dynamic optimization </li></ul>
  49. 49. Final Comments <ul><li>Biological Inspiration </li></ul><ul><ul><li>utility and extension </li></ul></ul><ul><ul><li>improved comprehension of natural phenomena </li></ul></ul><ul><li>Example based learning, where different pattern categories are represented by adaptive memories of the system </li></ul><ul><li>An iterative artificial immune network </li></ul>
  50. 50. <ul><li>CLONALG </li></ul><ul><ul><li>high degree of parallelism </li></ul></ul><ul><ul><li>by controlling the hipermutation rate an initial search for most general characteristics can be performed, followed by the search for smaller details </li></ul></ul><ul><ul><li>trade off between the clone size and the convergence speed </li></ul></ul><ul><ul><li>possibility of using heuristics to obtain global optima for problems like TSP </li></ul></ul>Final Comments
  51. 51. <ul><li>aiNet </li></ul><ul><ul><li>Models continuous spaces without the need of integration </li></ul></ul><ul><ul><li>Iterative model  dynamic models (DE) </li></ul></ul><ul><ul><li>Robustness with low redundancy </li></ul></ul><ul><ul><li>Clustering without a direct measure of distance* </li></ul></ul><ul><ul><li>RNA: knowledge distributed along the connections </li></ul></ul><ul><ul><li>aiNet: knowledge distributed in the cells </li></ul></ul><ul><ul><li> large amount of user defined parameters </li></ul></ul><ul><ul><li>Specific cells  general cells </li></ul></ul><ul><li>ABNET </li></ul><ul><ul><li>clustering, or grouping of similar patterns </li></ul></ul><ul><ul><li>capability of solving binary tasks </li></ul></ul>Final Comments
  52. 52. Final Comments <ul><li>Opt-aiNet </li></ul><ul><ul><li>;lsdkasdkj </li></ul></ul>
  53. 53. [email_address] Questions? Comments?

×