biologically-inspired computing lecture 18

Size: px

Start display at page:

Download "biologically-inspired computing lecture 18"

Julia Welch
5 years ago
Views:

1 Informatics -inspired lecture 18

Sections I485/H400 course outlook Assignments: 35% Students will complete 4/5 assignments based on algorithms presented in class Lab meets in I1 (West) 109 on Lab Wednesdays Lab 0 : January 14 th

2 Sections I485/H400 course outlook Assignments: 35% Students will complete 4/5 assignments based on algorithms presented in class Lab meets in I1 (West) 109 on Lab Wednesdays Lab 0 : January 14 th (completed) Introduction to Python (No Assignment) Lab 1 : January 28 th Measuring Information (Assignment 1) Graded Lab 2 : February 11 th L-Systems (Assignment 2) Graded Lab 3: March 25 th Cellular Automata & Boolean Networks (Assignment 3) Graded Lab 4: April 8 th Genetic Algorithms (Assignment 4) Due: April 22 nd Lab 5: April 22 nd Ant Clustering Algorithm (Assignment 5) Due May 4 th

3.10 Lecture notes Chapter 1: What is Life?

3 Readings until now Class Book Nunes de Castro, Leandro [2006]. Fundamentals of Natural Computing: Basic Concepts, Algorithms, and Applications. Chapman & Hall. Chapters 1, 2, 3, , , Lecture notes Chapter 1: What is Life? Chapter 2: The Logical Mechanisms of Life Chapter 3: Formalizing and Modeling the World Chapter 4: Self-Organization and Emergent Complex Behavior Chapter 5: Reality is Stranger than Fiction posted Other materials Flake s [1998], The Computational Beauty of Nature. MIT Press Chapters 20

final project schedule ALIFE 15 Projects

edu/alife14nyc/ 8 pages (LNCS proceedings

4 final project schedule ALIFE 15 Projects Due by May 4 th in Oncourse ALIFE 15 (14) Actual conference due date: pages (LNCS proceedings format) D= Preliminary ideas overdue! Individual or group With very definite tasks assigned per member of group

5 highlights Kyle Nealy Lab 3: CA & BN Matthew Remmel Darlan Farias

6 highlights Jonathan Stout Rafael Paiva Lab 3: CA & BN Lorander Saggu

The workings 1) Generate Random population of bit-strings 2) Evaluate Fitness Function for each decoded solution 3) Reproduce next generation Selection by fitness Variation crossover and mutation

7 The workings 1) Generate Random population of bit-strings 2) Evaluate Fitness Function for each decoded solution 3) Reproduce next generation Selection by fitness Variation crossover and mutation Fill new population 4) Go back to 2) until stop criteria is met Desired fitness Specified number of generations Convergence Lack of variability in population and/or fitness Tends to a peak f(x 3 ) genetic algorithms f(x i ) f(x 2 ) f(x 1 ) Parents

8 artificial genotype/phenotype mapping Search algorithms based on the mechanics of Natural Selection Based on distinction between a machine and a description of a machine Solution alternatives for optimization problems Genotype DNA Inherited variation computational evolution transcription translation (code) development RNA amino acid chains Traditional Genetic Algorithm Genotype environmental ramifications phenotype organism S S n 2 p!!! 1 S Variation Code: φ code x x x 1 2 n p Phenotype Selection

9 Types of encoding Binary encodings Typically fixed-length Many-letter encoding Larger alphabet (e.g. graph-generation grammars) Real-valued encodings Genes take real values Tree Encodings Genetic programming Indirect Encodings Modeling Phenotype development or posttranscription processes L-Systems, Dynamical systems, evolutionary robotics

homogenous lattice of state-determined systems t Cellular Automata x x-1 x+1 x t Possible neighborhood states K N Cellular automata Density Task

10 homogenous lattice of state-determined systems t Cellular Automata x x-1 x+1 x t Possible neighborhood states K N Cellular automata Density Task (a.k.a majority classification problem) #Lattices of 149 Binary Cells (599, 999) #Rules of Radius 3 (7 Cells in Neighborhood) #Task: Organize to < All 1's if Initial Configuration (IC) has more 1 Cells < All 0's if IC has more 0 Cells x x Possible CA transition functions K K N

encoding in GA with binary encoding t x x Possible neighborhood states Pop of rules 010010101100100 010010101100100 010010101100100 010010101100100 010010101100100 K N 2 7 128 Used in the

11 encoding in GA with binary encoding t x x Possible neighborhood states Pop of rules K N Used in the evolutionary search by GA (elite selection) code Traditional Genetic Algorithm Genotype S S 2 n p!!! 1 S Code: φ x1 x 2 x n p Phenotype Variation Selection Cellular automata

, Crutchfield,J.P., [1994]. "A genetic algorithm discovers particle-based computation in cellular automata".

12 With genetic algorithms Evolving CA rules #Das, Mitchell and Crutchfield < Used Genetic Algorithm to evolve rules for this task Typical Result: Block Expansion Regular domains {0+} {10+} {1+} Particles Das,R., Mitchell,M., Crutchfield,J.P., [1994]. "A genetic algorithm discovers particle-based computation in cellular automata". In: Parallel Problem Solving from Nature - PPSN III. Davidor,Y., Schwefel,H.-P., Manner,R. (Eds.), Springer-Verlag, pp

13 Informatics evolutionary algorithms to evolve photos with numerical encodings

In genetic algorithms real and integer encoding 1) Genotypes contain real or integer values 1) Crossover is performed in the same way 2) Mutation assigns a random number in a given interval 2) More

14 In genetic algorithms real and integer encoding 1) Genotypes contain real or integer values 1) Crossover is performed in the same way 2) Mutation assigns a random number in a given interval 2) More computationally demanding for Reals 3) Attention to crossover points 1) Conversion to binary avoids crossover issues, but longer genotypes x y r R G B n circles n Genes Agent Chromosome/Genotype (Population of p agents)

Tree encodings (no clear genotype) Fogel, Owens and Walsh (1966) evolving computer programs Artificial Intelligence through simulated evolution. Wiley.

15 Tree encodings (no clear genotype) Fogel, Owens and Walsh (1966) evolving computer programs Artificial Intelligence through simulated evolution. Wiley. Evolution of finite-state machines John Koza (1992) at Stanford University Genetic Programming: On the programming of computers by means of Natural Selection. MIT Press.

16 The workings 1) Generate Random population of tree/programs 2) Evaluate Fitness Function for each program Desired I/O, simplicity, speed 3) Reproduce next generation Selection by fitness Variation crossover and mutation Fill new population 4) Go back to 2) until stop criteria is met Desired fitness Specified number of generations Convergence + R * R + R+( *D) / f(x 3 ) / / / genetic programming / * / / D f(x 2 ) f(x 1 ) R R * - * C R*[(*C)- R] / + f(x i ) D * D R

17 A symbolic regression tool Eureqa Eureqa:

18 Gene expression programming Including a genotype/phenotype map in GP Proposed by Candida Ferreira Program trees are encoded in fixed-length linear genotypes Genotypes Open-reading frame architecture Stop signal not necessarily at end of genotype Non-coding genes are possible Can include genetic operators Genes contain two types of symbols Functions (only at the head) and terminals Multigenic solutions Assembled from non-coding operations between various open-reading frames C. FERREIRA [2001]. Gene Expression Programming: A New Adaptive Algorithm for Solving Problems. Complex Systems, 13 (2):

19 evolving morphologies and robots with indirect encodings Karl Sim s simulations and The Golem Project

20 objective function may be subjective evolutionary design "Once a Darwinian process gets going in a world, it has an open-ended power to generate surprising consequences: us, for example" Richard Dawkins Biomorphs

readings Next lectures Class Book Nunes de Castro, Leandro [2006]. Fundamentals of Natural Computing: Basic Concepts, Algorithms, and Applications. Chapman & Hall.

Chapter 2: The logical Mechanisms of Life Chapter 3: Formalizing and Modeling the World Chapter 4: Self-Organization and Emergent Complex Behavior Chapter 5: Reality is Stranger

21 readings Next lectures Class Book Nunes de Castro, Leandro [2006]. Fundamentals of Natural Computing: Basic Concepts, Algorithms, and Applications. Chapman & Hall. Chapter 3, all sections Sections 7.8 (evolving L-Systems), (biomorphs) Chapter 5, all sections Section 7.7, 8.3.1,8.3.6, Lecture notes Chapter 1: What is Life? Chapter 2: The logical Mechanisms of Life Chapter 3: Formalizing and Modeling the World Chapter 4: Self-Organization and Emergent Complex Behavior Chapter 5: Reality is Stranger than Fiction posted Optional materials Flake s [1998], The Computational Beauty of Life. MIT Press Chapter 20 Scientific American: Special Issue on the evolution of Evolution, January 2009.

Development. biologically-inspired computing. lecture 16. Informatics luis rocha x x x. Syntactic Operations. biologically Inspired computing

Development. biologically-inspired computing. lecture 16. Informatics luis rocha x x x. Syntactic Operations. biologically Inspired computing lecture 16 -inspired S S2 n p!!! 1 S Syntactic Operations al Code:N Development x x x 1 2 n p S Sections I485/H400 course outlook Assignments: 35% Students will complete 4/5 assignments based on algorithms