1 interactive evolutionary computation: a framework and applications november 10, 2009 sung-bae cho...

48
1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

Upload: elinor-copeland

Post on 20-Jan-2016

221 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

1

Interactive Evolutionary Computation:A Framework and Applications

November 10, 2009

Sung-Bae Cho

Dept. of Computer Science

Yonsei University

Page 2: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

2

Agenda

• Overview• Methodology

– IGA– Knowledge-based encoding– Partial evaluation based on clustering– Direct manipulation of evolution with NK-landscape model

• Applications– Media retrieval– Media design

• Concluding remarks

Page 3: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

3

User Modeling

AI

FL GA

NN

(Soft Computing)

Interaction/Evolution

Interactive Computational Intelligence

Overview

Human

Interface

Computer

(Emotion) (Efficiency)

Page 4: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

4

• Difficulty of optimization problems in interactive system– Outputs must be subjectively evaluated (graphics or music)

• Definition– Evolutionary computation that optimizes systems based on

subjective human evaluation– Fitness function is replaced by a human user

Interactive Evolutionary Computation (IEC)

Overview

Target systemf(p1,p2,…,pn)

InteractiveEC

System output

Subjective evaluationFrom Takagi (2001)

My goal is …

Page 5: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

5

Inherent Problems in IEC

• Human fatigue problem– Common to all human-machine interaction systems

• IEC– Acceleration of EC convergence with small population size and a

few generation numbers– Usually within 10 or 20 search generations

• Limitation of the individuals simultaneously displayed on a screen

• Limitation of the human capacity to memorize the time-sequentially displayed individuals

• Requirement to minimize human fatigue

Overview

I’m tired

Page 6: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

6

Proposed Framework

Overview

Conventional

IEC

Media Retrieval& Design

Proposed

IEC

Media Retrieval& Design

Domain Knowledge

PartialEvaluation

DirectManipulation

Page 7: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

7

Interactive Genetic Algorithm

Methodology

Crossover / Mutation Fitness Evaluation

Reproduction

Initial Population

Population

User Selection

GA IGA

Fitness Function

Page 8: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

8

Knowledge-based Encoding

Search Space

New Search Space

Domain Knowledge

Methodology

Removing impractical solutions

Using partially known information about the final solution

Focusing on a specific search space

Encoding in a modular manner

Effectiveencoding?

Page 9: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

9

Partial Evaluation based on Clustering

Methodology

Issues Clustering algorithm selection, Indirect fitness allocation strategy

Page 10: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

10

Direct Manipulation of Evolution

Methodology

IGA

Direct Manipulation

It is very similar to the final

solution but a bit different.

Proof of usefulness of the direct manipulation NK-landscape

Page 11: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

11

Overview

Applications

Media Retrieval Media Design

Image Retrieval

Video Retrieval

Music Retrieval

Fashion Design

Flower Design

Intrusion PatternDesign

Humanized Media Retrieval & Design

Page 12: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

12

Content-based Image Retrieval

• Keyword-based method– Much time and labor to construct indexes in a large databases– Performance decrease when index constructor and user have

different point of view– Inherent difficulty to describe image as a keyword

• Content-based method– Image contents as a set of features extracted from image– Specifying queries using the features

Applications: Image Retrieval

Page 13: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

13

Motivation

• Systems for content-based image retrieval– QBIC (IBM, 1993)– QVE (Hirata and Kato, 1993)– Chabot (Berkeley, 1994)

• Needs for image retrieval based on human intuition– Difficult to get perfect expression by queries– Lack of expression capability

Applications: Image Retrieval

Page 14: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

14

System Structure

Applications: Image Retrieval

Page 15: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

15

Chromosome Structure

Applications: Image Retrieval

Original image Transformed image Chromosome

0 1 2 3 49 50

R

G

B

Page 16: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

16

Convergence Test

• The case of gloomy image retrieval

Applications: Image Retrieval

Initial population

The 8th population

Page 17: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

17

Subjective Test by Sheffe’s Method

Applications: Image Retrieval

Confidence interval 95% 99%

Page 18: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

18

Motivation

• Evolution of computing technology (Huge computing storage, networking)– Necessary for management of video data transfer, processing

and retrieval

• Change of view point– Query by keyword Query by content Difficult to represent

human’s semantic information– Query by semantics

• Emotion-based retrieval

Applications: Video Retrieval

Page 19: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

19

Hierarchical Structure of Video

Applications: Video Retrieval

Page 20: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

20

System Architecture

Applications: Video Retrieval

Page 21: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

21

Chromosome Structure

Applications: Video Retrieval

Page 22: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

22

System Interface

Applications: Video Retrieval

Page 23: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

23

User’s Satisfaction

Applications: Video Retrieval

Page 24: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

24

Emotional Music Retrieval

• Music information retrieval Immature field• Query by singing• Query by content

– Uploading pre-recording humming via web browser– Typing the string to represent the melody contour

• Preprocessing– Query by humming transforms user’s humming into

symbolized representation• Conventional music retrieval system

– Fast and effective extraction of musical patterns and transformation of user’s humming into systematic notes

• Problem : User cannot remember some parts of melody

Applications: Music Retrieval

Page 25: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

25

Power Spectrum Analysis

Applications: Music Retrieval

b

a

c

d

1/f

b

a

c

d

1/f

b

a

c

d

1/f

)](

[lo

g 10f

S Z

)(log10 f

a : classicb : jazzc : rockd : news channel

Page 26: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

26

System Architecture

Applications: Music Retrieval

MusicDatabase

:

③ Retrieved music

② Retrieval of music by the query

IGA

① Query

⑤ Generation of new offspring

④ User evaluates the music retrieved

MusicDatabase

:

③ Retrieved music

② Retrieval of music by the query

IGA

① Query

⑤ Generation of new offspring

④ User evaluates the music retrieved

Page 27: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

27

Chromosome Structure

Applications: Music Retrieval

Page 28: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

28

Convergence Test (1)

Applications: Music Retrieval

0

10

20

30

40

50

60

0 1 2 3 4 5 6 7 8 9 10

generation

Dis

tance

③ m=0.5 c=0.2

① m=0.5 c=0.5

② m=0.2 c=0.5

④ m=0.2 c=0.2

Page 29: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

29

Convergence Test (2)

Applications: Music Retrieval

Page 30: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

30

Change of Consumer Economy

Applications : Fashion Design

Before the Industrial Revolution : Customers have few choices on buying their clothes

After the Industrial Revolution :Customers can make their choices with very large variety

Near Future :Customers can order and get clothes of their favorite design

ManufacturerOriented

ConsumerOriented

Page 31: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

31

Need for Interactive System

• Almost all consumers are non-professional at design

• To make designers contact all consumers is not effective

• Need for the design system that can be used by non-professionals

Applications : Fashion Design

Page 32: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

32

Fashion Design

• Definition – To make a choice within various

styles that clothes can take

• Three shape parts of fashion design– Silhouette– Detail– Trimming

Applications : Fashion Design

Page 33: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

33

System Architecture

Decode

User Fitness

Combine

Display

Interactive Genetic Algorithm

OpenGL Program

GA operation Reproduce

Models ofeach part

Applications : Fashion Design

Page 34: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

34

Knowledge-based Encoding

Search space size=34*8*11*8*9*8=1,880,064

A : Neck and body style(34) E : Skirt and waistline style(9)

C : Arm and sleeve style(11)

B : Color(8)

D : Color(8)

……

A B C D E F

Total 23 bits

F : Color(8)

Applications : Fashion Design

Page 35: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

35

Direct Manipulation Interface

Applications : Fashion Design

Page 36: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

36

Fitness Changes for Encoding Schemes

Applications : Fashion Design

0

10

20

30

40

50

60

70

1 2 3 4 5 6 7 8 9 10

Generation

Ave

rage

Fitn

ess

Knowledge- based Encoding

Sequential Encoding

Page 37: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

37

Hypercube Analysis

Applications : Fashion Design

0000 bishop

0001 china

1000 poncho

0010 flare

0100 mandarin

0011 french

1001 tight

0101 melon

0111 pagoda

1010 tucked

0110 mutton

1110

No mapped

design

1011 no sleeves

1111

No mapped

design

1101

No mapped

design

1100

No mapped

design

65

45

20

80

8535

80

50

15

6075

100

0000 bishop

0001 china

1000 poncho

0010 flare

0100 mandarin

0011 french

1001 tight

0101 melon

0111 pagoda

1010 tucked

0110 mutton

1110

No mapped

design

1011 no sleeves

1111

No mapped

design

1101

No mapped

design

1100

No mapped

design

DM

65

45

20

80

8535

80

50

15

6075

100

One-Mutant Neighbors using Genetic Operator Shortest Path Using DM Method

Page 38: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

38

Motivation

• Proposed Method– Representation of knowledge-based genotype using the

structure of real flower– Automatic generation of creatures

• Process– Directed graph: Define the genotype of flower structure– IGA: Evaluation of created individuals and automatic creation– Math Engine: Representation of phenotype

• Objective– Creation of character morphology similar to real flower– Find optimal solution in small solution space using knowledge-

based genotype representation such as directed graph– Automatic creation of character using IGA

Applications : Flower Design

Page 39: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

39

Related Works

• Tree, flower and feather modeling– L-System

• Box-based artificial characters (Karl Sims) evolution– Graph model representation

• Golem– Robotic form and controller

evolution– Graph-based data structure

Applications : Flower Design

Page 40: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

40

Overall Procedure

Applications : Flower Design

Create initial edges between nodes

Set dimension and color of each part

Set parameters of parts and edges

Create rigid parts in genotype

Create 3D world in simulation

Perform 3D rendering

Evaluate fitness of each individual

Satisfied individual ?

Selection

Crossover/Mutation

Create individuals of next generation

Stop

YesNo

Page 41: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

41

Chromosome Structure

Applications : Flower Design

Page 42: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

42

Convergence Test

Applications : Flower Design

0

2

4

6

8

10

1 2 3 4 5 6 7 8 9 10

Generation

Fitn

ess

valu

e

Average fitness value Best fitness value

Page 43: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

43

Subjective Test by Sheffe’s Method

Applications : Flower Design

- 3 - 2 - 1 0 1 2 3

99%

95%

Page 44: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

44

Motivation (1)

• Various intrusion detection systems– For evaluating, IDS uses known intrusions

• Possibility of having different patterns within identical intrusion type

• Difficult to detect transformed intrusions

• Problem– For better evaluation, intrusion patterns are needed more

• Difficult to generate all possible proper intrusion patterns manually

– Automatic generation is needed

Applications : Intrusion Pattern Design

Page 45: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

45

Motivation (2)

• Evolutionary pattern generation– Needs a variety of intrusion patterns

• It is possible to generate new patterns by combining primitives

– Generation of transformed intrusion patterns using simple genetic algorithm

• IGA-based intrusion pattern generation– Difficult to estimate the fitness of patterns automatically– Using interactive genetic algorithm for the evaluation of patterns

Applications : Intrusion Pattern Design

Page 46: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

46

Method

• Intrusive behavior can be conducted in 5 steps– Proposed by DARPA– Possible path of U2R attacks

web download

editor

ftp

mail

floopy

archive

octal character

simpleencryption

characterstuffing

encoding

shell variables

shell script

encryptedshell interaction

generate chaftoutput in bg

delaytedexecution

change filepermissions

display files

delete files

transfer files

alter files

remove files

restorepermissions

edit audit logs

TRANSPORT ENCODING EXECUTION ACTIONS CLEAN UP

Applications : Intrusion Pattern Design

Page 47: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

47

System Architecture

Initial Population

KnownIntrusion Patterns

IGA

GA

FitnessEvaluation

ExploitCode

Intrusion Patterns

GenerateAudit Data

000100.....001100.....

Applications : Intrusion Pattern Design

Page 48: 1 Interactive Evolutionary Computation: A Framework and Applications November 10, 2009 Sung-Bae Cho Dept. of Computer Science Yonsei University

48

Concluding Remarks

• Humanized CI framework using IEC– Knowledge-based encoding– Partial evaluation based on clustering– Direct manipulation of evolution with NK-landscape

• Applications in media retrieval & design– Image, video and music retrievals– Fashion, flower and intrusion pattern designs

• Contribution of humanized CI to engineering fields

Developing Emotional HC Interface by Evolving models through Interaction with Humans