Computers and Creativity (Book)

This book is a study of human creative behaviour from a computational modelling perspective. The authors examine theories and models of the creative process in humans, both input creativity - the scientific, analytic side of devising interpretations of input information - and output creativity - the artistic, synthetic process of generating something novel and innovative. 

After a critical examination of both earlier theories and computational models, the authors develop and then formally define their own model, an emergent-memory theory. This theory is implemented (in several variants) as a computational model and a detailed empirical study is reported and analysed. 

The final chapter is an extensive discussion of the conclusions to be drawn both with respect to their emergent-memory theory and to fundamental issues of creativity that were raised in the beginning of the book. 

Edition

1  Introduction

1.1  The nature of creativity ......................................... 1 

1.2  Aspects of creativity .............................................. 2 

1.3  Social aspects of creative behaviour ........................ 6

1.4  Some definitions ................................................... 7

1.5  The art and the science .......................................... 8

2  Theories of creativity 

2.1  Poincaré and the four-stage model .......................... 10

2.2  Weisberg's criticism of the four-stage model ............. 11

2.3  A critique of Weisberg's view .................................. 15

2.4  The cortical-arousal theory of creativity ................... 18 

2.5  Some characteristics of creativity ............................ 21 

3  Computational modelling

3.1  The computational modelling approach .................... 24

3.2  The benefits and drawbacks ................................... 26

3.3  Previous computational models .............................. 28

3.3.1  Randomness and rigidity .............................. 29

3.3.2  Generative grammars .................................. 30

3.3.3  Discovery programs .................................... 33

3.3.4  Meta-rules ................................................. 42 

3.3.5  Analogical mechanisms ................................ 44 

3.3.6  Towards flexible representations .................... 46

3.3.7  Decentralised systems .................................. 49

3.3.8  The neural-network issue .............................. 52 

4  Emergent-memory models 

4.1  A characteristic of computational modelling .............. 56 

4.2  A theory of emergent memory ................................ 58

4.3  A formal definition ................................................. 59

4.3.1  Builders ....................................................... 60

4.3.2  K-lines ......................................................... 61 

4.3.3  Censors and suppression ................................ 65 

4.3.4  Credit assignment ......................................... 68 

4.3.5  Task selection ............................................... 69 

5  A Computation Model: GENESIS 

5.1  Computational decisions ......................................... 74 

5.2  Micro-Eleusis as a domain ....................................... 76 

5.3  Domain-specific details ........................................... 79 

5.4  Implementation details ........................................... 87 

5.5  Alternative computational models ............................ 90

5.5.1  SPARC/μE .................................................... 91

5.5.2  A genetic representation ................................ 91

5.5.3  Two classifier systems ................................... 92

5.5.4  The NODDY algorithm ................................... 92

6  Empirical Studies 

6.1  The scope of the empirical study ............................. 94

6.2  The 24 sequences ................................................. 96

6.3  GENESIS at work ................................................. 100

6.3.1  Best-first behaviour ..................................... 100

6.3.2  Terraced-scan behaviour .............................. 113

6.4  Sequence analysis ............................................... 119 

6.4.1  GENESIS-BF .............................................. 119

6.4.2  GENESIS-TS .............................................. 121

6.4.3  The classifier systems ................................. 124 

6.4.4  SPARC/μE's analyses .................................. 125

6.5  Programs as dealers ............................................ 125 

6.6  Programs as players ............................................ 129 

6.7  Playing with people ............................................. 131     

7  Conclusions To Be Drawn 

7.1  Theoretical results .............................................. 140

7.2  Results from the Implementation Stage ................. 141

7.3  Experimental Results .......................................... 142 

7.4  Empirical Conclusions ......................................... 144

7.5  The "aha!" phenomenon: cause or effect? ............. 145

7.6  Chunking and magic numbers ............................. 146 

7.7  Controlled randomness ...................................... 147

7.8  Breaking out and going wrong ............................ 150

7.9  Conceptual spaces ............................................ 154 

7.10  Mere novelty real creativity .............................. 157 

7.11  Summary points ............................................. 160 

A  The Twenty-four Test Sequences                                    169  

 

 

Related Titles