Archaeology and the application of artificial intelligence : case-studies on
use-wear analysis of prehistoric flint tools
Dries, M.H. van den
Citation
Dries, M. H. van den. (1998, January 21). Archaeology and the application of artificial intelligence :
case-studies on use-wear analysis of prehistoric flint tools. Retrieved from
https://hdl.handle.net/1887/13148
Version:
Corrected Publisher’s Version
License:
Licence agreement concerning inclusion of doctoral thesis in the Institutional
Repository of the University of Leiden
Downloaded from:
https://hdl.handle.net/1887/13148
Archaeological Studies Leiden University
Archaeology and the application of
Artificial Intelligence
Case-studies on use-wear analysis
of prehistoric flint tools
Faculty of Archaeology University of Leiden 1998
Proefschrift
ter verkrijging van de graad van Doctor
aan de Rijksuniversiteit te Leiden,
op gezag van de Rector Magnificus Dr. W.A. Wagenaar,
hoogleraar in de Faculteit der Sociale Wetenschappen,
volgens besluit van het College van Dekanen
te verdedigen op 21 januari 1998
te klokke 15.15 uur
door
Monique Henriëtte
VAN DEND
RIESPromotiecommissie
promotor: Prof.dr. L.P. Louwe Kooijmans
co-promotores: Dr. A.L. van Gijn Dr. H. Kamermans
referent: Prof.dr. J.E. Doran (University of Essex, Colchester)
1 Introduction 9
2 The application of artificial intelligence in archaeology 11 2.1 Introduction 11
2.2 Historical and contextual background 11 2.2.1 The emergence of quantitative methods 11 2.2.2 A new approach: artificial intelligence 12
2.3 Review of knowledge-based applications in archaeology 13 2.4 Attitudes towards the application of knowledge-based methods 15 2.4.1 From high expectations to cautiousness 15
2.4.2 Lessons 15 2.4.3 Discussion 16
3 Expert system fundamentals 19 3.1 Introduction 19
3.2 Architecture 19
3.3 Knowledge representation and reasoning methods 21 3.3.1 Introduction 21 3.3.2 Predicate logic 21 3.3.3 Production rules 23 3.3.4 Semantic nets 24 3.3.5 Frames 25 3.3.6 Hybrid representation 26
3.4 Expert-system development process 26 3.4.1 Introduction 26
3.4.2 Orientation 26
3.4.3 Knowledge acquisition 27
3.4.4 Design, implementation and evaluation 28 3.5 Implementation tools 28
3.5.1 Introduction 28
3.5.2 Programming languages versus shells 29
4 The application domain: use-wear analysis of prehistoric flint tools 31 4.1 Introduction 31
4.2 The emergence and development of use-wear analysis 31 4.3 Methodical aspects 33
4.3.1 Combination of information sources 33 4.3.2 Use-wear phenomena 33
4.3.3 Microscopy 35
4.4 Difficulties encountered 36
4.5 From a qualitative to a quantitative method? 37 4.5.1 Introduction 37
5
4.5.2 Automation of the observation process 38 4.5.3 Quantification of recordings 40
4.5.4 Automation of the inferencing process 40 4.6 Discussion 41
5 An expert system application for use-wear analysis: WAVES 43 5.1 Introduction 43
part one: WAVES under construction 44 5.2 The knowledge acquisition 44
5.2.1 Introduction 44
5.2.2 The elicitation of the expert knowledge 44 5.2.3 The data analyzed 46
5.2.4 Knowledge modelling and uncertainty handling 49 5.3 Design and implementation 51
5.3.1 Introduction 51 5.3.2 Knowledge handling 51 5.3.2.1 Requirements 51 5.3.2.2 Decisions 52 5.3.3 User interface 53 5.3.3.1 Requirements 53 5.3.3.2 Decisions 53
5.3.4 Hardware and software 54 5.3.4.1 Requirements 54
5.3.4.2 Decisions 54 5.4 Discussion 55 5.4.1 Introduction 55
5.4.2 On using knowledge derived from experiments 56 5.4.3 On managing uncertainty 56
part two: WAVES in action 57
5.5 The composition of the analysis procedure 57
5.6 The composition of the hypothesis validation procedure 60 5.7 A session 61
5.7.1 Getting started 61
5.7.2 Running the analysis procedure 63 5.7.3 Interpreting the interpretation 66
5.7.4 Running the hypothesis validation procedure 69 5.8 An assessment of the composition of the application 72 5.8.1 Introduction 72
5.8.2 Answers to expectations of computer archaeologists 72 5.8.3 Answers to expectations of use-wear analysts 74 5.8.4 Unanswered questions and suggestions for additions 76 5.9 Comparison with FAST 77
6 A neural network prototype for use-wear analysis: WARP 79 6.1 Introduction 79
6.2 Neural network fundamentals 80 6.2.1 Historical backgrounds 80 6.2.2 Architecture 81
6.3 Development process 83 6.3.1 Introduction 83
6.3.2 Knowledge representation 83
6.3.3 A matter of training and experimentation 84
6.3.4 A session 85 6.4 The prototype 86 6.4.1 Introduction 86
6.4.2 The knowledge represented 86 6.4.3 The training process 88
6.5 An assessment of the prototype 89 6.5.1 Introduction 89
6.5.2 Answers to expectations of computer archaeologists 89 6.5.3 Answers to expectations of use-wear analysts 90 6.5.4 Suggestions for additions 91
7 Test results 93 7.1 Introduction 93
7.2 Blind tests in use-wear analysis 94 7.3 The first test 95
7.3.1 Test-set composition 95
7.3.2 The expert system’s achievements 96 7.3.3 The neural network’s achievements 97 7.3.4 Conclusion 98
7.4 The second test 99 7.4.1 Introduction 99 7.4.2 Test-set composition 99 7.4.3 Achievements 105 7.4.4 Conclusion 113
7.4.5 Comparison with other blind tests 117 7.5 Discussion 120
8 Synthesis 123 8.1 Introduction 123
8.2 The added value of expert systems and neural networks 123 8.2.1 Expert systems 123
8.2.2 Neural networks 124
8.2.3 Expert systems versus neural networks? 125
8.3 Added value of knowledge-based systems for archaeology 125 8.4 Recommendations 126
8.4.1 A different approach of functionality 126 8.4.2 Care for social acceptability 127
8.4.3 Change of attitude 128 8.5 Concluding remarks 129
Appendix I: The conceptual knowledge of the analysis procedure 131
Appendix II: The conceptual knowledge of the hypothesis validation procedure of WAVES 147
Appendix III: The input and output variables of WARP 153
Appendix IV: Descriptions of the blind test 155
Appendix V: Interpretations of the blind test 171
References 187
Samenvatting (Dutch summary) 195
Acknowledgements 201
Addendum: A.L. van Gijn, Richard Fullagar: Welcoming Waves 203