Hybrid Evolution of Heterogeneous Neural Networks

Similar documents
Integer Programming: Algorithms - 3

Comparing Neural Networks and ARMA Models in Artificial Stock Market

Automatic parameter regulation for a tracking system with an auto-critical function

International Journal of Software and Web Sciences (IJSWS)

CHAPTER 3 SECURITY CONSTRAINED OPTIMAL SHORT-TERM HYDROTHERMAL SCHEDULING

Overview. Swarms in nature. Fish, birds, ants, termites, Introduction to swarm intelligence principles Particle Swarm Optimization (PSO)

Genetic Algorithm. Based on Darwinian Paradigm. Intrinsically a robust search and optimization mechanism. Conceptual Algorithm

Load balancing in a heterogeneous computer system by self-organizing Kohonen network

College of information technology Department of software

Memory Allocation Technique for Segregated Free List Based on Genetic Algorithm

Web Service Selection using Particle Swarm Optimization and Genetic Algorithms

Evaluation of Crossover Operator Performance in Genetic Algorithms With Binary Representation

Pattern recognition using multilayer neural-genetic algorithm

Research on the Performance Optimization of Hadoop in Big Data Environment

ISSN: ISO 9001:2008 Certified International Journal of Engineering Science and Innovative Technology (IJESIT) Volume 2, Issue 3, May 2013

DECISION TREE INDUCTION FOR FINANCIAL FRAUD DETECTION USING ENSEMBLE LEARNING TECHNIQUES

Alpha Cut based Novel Selection for Genetic Algorithm

Feature Selection for Stock Market Analysis

About the Author. The Role of Artificial Intelligence in Software Engineering. Brief History of AI. Introduction 2/27/2013

Cellular Automaton: The Roulette Wheel and the Landscape Effect

Multiple Layer Perceptron Training Using Genetic Algorithms

A HYBRID GENETIC ALGORITHM FOR THE MAXIMUM LIKELIHOOD ESTIMATION OF MODELS WITH MULTIPLE EQUILIBRIA: A FIRST REPORT

A Review And Evaluations Of Shortest Path Algorithms

A Review of Data Mining Techniques

Keywords revenue management, yield management, genetic algorithm, airline reservation

Machine Learning and Data Analysis overview. Department of Cybernetics, Czech Technical University in Prague.

A Novel Binary Particle Swarm Optimization

PERFORMANCE ANALYSIS OF HYBRID FORECASTING MODEL IN STOCK MARKET FORECASTING

Original Article Efficient Genetic Algorithm on Linear Programming Problem for Fittest Chromosomes

Architectural Design for Space Layout by Genetic Algorithms

Flexible Neural Trees Ensemble for Stock Index Modeling

PLAANN as a Classification Tool for Customer Intelligence in Banking

A Simultaneous Solution for General Linear Equations on a Ring or Hierarchical Cluster

Evolutionary SAT Solver (ESS)

Filling the Semantic Gap: A Genetic Programming Framework for Content-Based Image Retrieval

CLOUD DATABASE ROUTE SCHEDULING USING COMBANATION OF PARTICLE SWARM OPTIMIZATION AND GENETIC ALGORITHM

New binary representation in Genetic Algorithms for solving TSP by mapping permutations to a list of ordered numbers

Stock price prediction using genetic algorithms and evolution strategies

Asexual Versus Sexual Reproduction in Genetic Algorithms 1

A Brief Study of the Nurse Scheduling Problem (NSP)

Feature Selection using Integer and Binary coded Genetic Algorithm to improve the performance of SVM Classifier

The Applications of Genetic Algorithms in Stock Market Data Mining Optimisation

Estimation of the COCOMO Model Parameters Using Genetic Algorithms for NASA Software Projects

Introduction to Machine Learning and Data Mining. Prof. Dr. Igor Trajkovski

Optimal PID Controller Design for AVR System

Genetic algorithms for changing environments

Czech Technical University in Prague Faculty of Electrical Engineering. using. Pavel Kordík

Practical Applications of Evolutionary Computation to Financial Engineering

Vol. 35, No. 3, Sept 30,2000 ملخص تعتبر الخوارزمات الجينية واحدة من أفضل طرق البحث من ناحية األداء. فبالرغم من أن استخدام هذه الطريقة ال يعطي الحل

Model-based Parameter Optimization of an Engine Control Unit using Genetic Algorithms

Genetic algorithm evolved agent-based equity trading using Technical Analysis and the Capital Asset Pricing Model

A Hybrid Tabu Search Method for Assembly Line Balancing

STA 4273H: Statistical Machine Learning

Research Article Service Composition Optimization Using Differential Evolution and Opposition-based Learning

An Overview of Knowledge Discovery Database and Data mining Techniques

Performance Optimization of I-4 I 4 Gasoline Engine with Variable Valve Timing Using WAVE/iSIGHT

A Basic Guide to Modeling Techniques for All Direct Marketing Challenges

Internal Evolution for Agent Cognition Agent-Based Modelling of an Artificial Stock Market

Numerical Research on Distributed Genetic Algorithm with Redundant

Constrained Classification of Large Imbalanced Data by Logistic Regression and Genetic Algorithm

NEURAL NETWORK FUNDAMENTALS WITH GRAPHS, ALGORITHMS, AND APPLICATIONS

GARDA: a Diagnostic ATPG for Large Synchronous Sequential Circuits

Hybrid processing of SCADA and synchronized phasor measurements for tracking network state

Course DSS. Business Intelligence: Data Warehousing, Data Acquisition, Data Mining, Business Analytics, and Visualization

Genetic Algorithm Based Interconnection Network Topology Optimization Analysis

THREE DIMENSIONAL REPRESENTATION OF AMINO ACID CHARAC- TERISTICS

A hybrid Approach of Genetic Algorithm and Particle Swarm Technique to Software Test Case Generation

Introduction To Genetic Algorithms

A Genetic Programming Model for S&P 500 Stock Market Prediction

2: Computer Performance

SELECTING NEURAL NETWORK ARCHITECTURE FOR INVESTMENT PROFITABILITY PREDICTIONS

Neural Network and Genetic Algorithm Based Trading Systems. Donn S. Fishbein, MD, PhD Neuroquant.com

Volume 3, Issue 2, February 2015 International Journal of Advance Research in Computer Science and Management Studies

Chapter 5 Business Intelligence: Data Warehousing, Data Acquisition, Data Mining, Business Analytics, and Visualization

A Genetic Algorithm Processor Based on Redundant Binary Numbers (GAPBRBN)

Lecture. Simulation and optimization

Correspondence should be addressed to Chandra Shekhar Yadav;

Solving Method for a Class of Bilevel Linear Programming based on Genetic Algorithms

Un algorithme génétique hybride à gestion adaptative de diversité pour le problème de tournées de véhicules et ses variantes

Evolutionary Algorithms using Evolutionary Algorithms

Incremental Evolution of Complex Light Switching Behavior

Online Evolution of Deep Convolutional Network for Vision-Based Reinforcement Learning

Design of Demand Forecasting Expert System for Dynamic Supply Chains

How To Find Out If Novelty Search Is A Good Way To Find A Novel Solution

Predict Influencers in the Social Network

Data Mining Applications in Manufacturing

Effect of Using Neural Networks in GA-Based School Timetabling

Transcription:

Hybrid Evolution of Heterogeneous Neural Networks 01001110 01100101 01110101 01110010 01101111 01101110 01101111 01110110 01100001 00100000 01110011 01101011 01110101 01110000 01101001 01101110 01100001 00100000 01101011 01100001 01110100 01100101 01100100 01110010 01111001 00100000 01110000 01101111 01100011 01101001 01110100 01100001 01100011 01110101 00101100 00100000 01000110 01000101 01001100 00100000 01000011 01010110 01010101 01010100 Zdeněk Buk bukz1@fel.cvut.cz Miroslav Šnorek snorek@fel.cvut.cz http://cig.felk.cvut.cz Computational Intelligence Group Department of Computer Science and Engineering Faculty of Electrical Engineering Czech Technical University in Prague ICANN 2008

Outline Continual Evolution Algorithm (CEA) description Data structures encoding of the individuals Evolution process Control functions Testing, experiments Population behavior Conclusion

Continual Evolution Algorithm Hybrid genetic algorithm combination of genetic and gradient-based methods Separate evolution of structure and parameters of individuals (models, neural networks) Variable population size Sequential replacement of individuals evolution in continual time dimension age parameter of each individual

Individuals encoding x i = a i, p i, s i, b i Separate evolution of structure and parameters of individuals (models, neural networks)

Individuals encoding x i = a i, p i, s i, b i age of the individual

Individuals encoding x i = a i, p i, s i, b i structural vector

Individuals encoding x i = a i, p i, s i, b i structural vector topology of the network

Individuals encoding x i = a i, p i, s i, b i structural vector topology of the network activation functions

Individuals encoding x i = a i, p i, s i, b i parametric vector behavioral vector

Individuals encoding x i = a i, p i, s i, b i parametric vector behavioral vector weights

Individuals encoding x i = a i, p i, s i, b i

Evolution process Hybrid genetic algorithm combination of genetic and gradient-based methods 2 Dimensional evolution

Evolution process

Crossover and mutation Individual 1 p 1 s 1 b 1 p 2 s 2 b 2 Offspring Individual 2

Crossover and mutation Individual 1 p 1 s 1 b 1 p 2 s 2 b 2 Random mutation Offspring Individual 2

Crossover and mutation Individual 1 p 1 s 1 b 1 p 2 s 2 b 2 Random mutation Offspring Individual 2

Crossover and mutation Individual 1 p 1 s 1 b 1 p 2 s 2 b 2 Offspring Individual 2

Crossover and mutation Individual 1 p 1 s 1 b 1 p 2 s 2 b 2 Offspring Individual 2

Crossover and mutation Individual 1 p 1 s 1 b 1 p 2 s 2 b 2 Offspring Individual 2 Random mutation

Crossover and mutation Individual 1 p 1 s 1 b 1 p 2 s 2 b 2 Offspring Copy Individual 2

Crossover and mutation Individual 1 p 1 s 1 b 1 p 2 s 2 b 2 Offspring p i s i b i Individual 2

Time dimension Training the behavioral vector during time using gradient algorithm. Age=0 p i s i b i Time

Time dimension Training the behavioral vector during time using gradient algorithm. Age=0 Age=1 p i s i b i Time

Time dimension Training the behavioral vector during time using gradient algorithm. Age=0 Age=1 Age=2 p i s i b i Time

Probability control Probability functions RP reproduction probability, DP death probability, elimination of bad solutions in population. Balancing functions to keep the population size in some reasonable limits.

Reproduction probability Raw reproduction probability function * * RP x i = RP a i, F x i Fitness RP* Age Fitness Age

Death probability Raw death probability function * * DP x i = DP a i, F x i Fitness DP* Age Fitness Age

Balancing functions Computation of final probabilities. Depend on raw probabilities and population size. Big population grows slower, small population grows faster. * RP x i = BAL RP N, RP x i * DP x i = BAL DP N, DP x i

Evolution control

Evolution control

Evolution control

Evolution control

Evolution control

Evolution control

Evolution control

Testing, experiments Construction of the neural networks. Universal topology based on fully recurrent network. Structure topology adjacency matrix activation functions parametrized Λ-functions Behavior weight matrix

Testing, experiments Neural networks based on fully recurrent topology construction using CEA Activation function optimization

Testing, experiments Benchmark tasks Learn to oscillate experiment

Population behavior

Population behavior

Fitness Population behavior Iterations

Conclusion Reduction of the number of fitness function evaluations. Using the probability and balancing functions it is possible to change the CEA to behave more like random search, standard genetic algorithm, or gradient algorithm.

Conclusion CEA universal optimization algorithm mainly for problems with separate description of structure and behavior typically neural networks. Automatic control of the size of the population exploitation exploration