Uncertainty in Artificial Intelligence 2 (eBook)

L.N. Kanal, J.F. Lemmer (Herausgeber)

eBook Download: PDF

2014 | 1. Auflage
469 Seiten
Elsevier Science (Verlag)
978-1-4832-9653-1 (ISBN)

Infos

This second volume is arranged in four sections: Analysis contains papers which compare the attributes of various approaches to uncertainty. Tools provides sufficient information for the reader to implement uncertainty calculations. Papers in the Theory section explain various approaches to uncertainty. The Applications section describes the difficulties involved in, and the results produced by, incorporating uncertainty into actual systems.

Inhalt

Front Cover 1
Uncertainty in Artificial Intelligence 2 4
Copyright Page 5
Table of Contents 10
PREFACE 6
CONTRIBUTORS 8
PART I. ANALYSIS 14
CHAPTER 1. MODELS VS. INDUCTIVE INFERENCE FOR DEALING WITH PROBABILISTIC KNOWLEDGE 16
1. Introduction 16
2. Structures on Event Spaces 16
3. Webs 18
4. Induction and maximum-entropy 19
5. Discussion 20
References 21
CHAPTER 2. AN AXIOMATIC FRAMEWORK FOR BELIEF UPDATES 24
1. INTRODUCTION 24
2. SCOPE OF THE AXIOMIZATION 25
3. FUNDAMENTAL PROPERTIES FOR A MEASURE OF ABSOLUTE BELIEF 25
4. FUNDAMENTAL PROPERTIES FOR A MEASURE OF CHANGE IN BELIEF 27
5. A CONSEQUENCE OF THE AXIOMS 28
6. PROBABILISTIC BELIEF UPDATES 30
7. CONCLUSIONS 33
ACKNOWLEDGEMENTS 33
NOTES 33
CHAPTER 3. THE MYTH OF MODULARITY IN RULE-BASED SYSTEMS FOR REASONING WITH UNCERTAINTY 36
1. INTRODUCTION 36
2. OVERVIEW OF THE MYCIN CERTAINTY FACTOR MODEL 38
3. DEFINITION OF SEMANTIC MODULARITY 38
4. CONSEQUENCES OF MODULARITY 39
5. THE MYTH OF MODULARITY 43
6. A WEAKER NOTION OF MODULARITY 43
7. SUMMARY 46
ACKNOWLEDGEMENTS 46
NOTES 46
REFERENCES 47
CHAPTER 4. IMPRECISE MEANINGS AS A CAUSE OF UNCERTAINTY IN MEDICAL KNOWLEDGE-BASED SYSTEMS 48
1. INTRODUCTION: LEXICAL IMPRECISION 48
2. EFFECTS ON KNOWLEDGE-BASED SYSTEMS 50
3. LEXICAL IMPRECISION IS NOT LEXICAL AMBIGUITY 51
4. COPING WITH LEXICAL IMPRECISION 52
5. SUMMARY 53
ACKNOWLEDGMENTS 54
REFERENCES 54
CHAPTER 5. Evidence as Opinions of Experts 56
Abstract 56
1. Introduction 56
2. The Rule of Combination and Normalization 57
3. Spaces of Opinions of Experts 58
4. Equivalence with the Dempster/Shafer Rule of Combination 62
5. An Alternative Method for Combining Evidence 62
6. Conclusions 65
Acknowledgements 65
References 66
CHAPTER 6. PROBABILISTIC LOGIC: SOME COMMENTS AND POSSIBLE USE FOR NONMONOTONIC REASONING 68
Introduction 68
Probabilistic Logic 68
Nonmonotonic Reasoning 71
An Inconsistent System 71
Discussion of the Entailment Results 72
Some Computational Observations 74
Conclusions 74
References: 75
CHAPTER 7. EXPERIMENTS WITH INTERVAL-VALUED UNCERTAINTY 76
1. INTRODUCTION 76
2. THE INFORMATION RETRIEVAL MODEL 76
3. MODELS OF UNCERTAINTY 77
4. THE INFERENCE ENGINE 79
5. EXPERIMENTS 80
6. DISCUSSION 86
REFERENCES 88
CHAPTER 8. EVALUATION OF UNCERTAIN INFERENCE MODELS I: PROSPECTOR 90
1. INTRODUCTION 90
2. OVERVIEW OF THE PROSPECTOR MODEL 91
3. METHOD 92
4. RESULTS 94
5. DISCUSSION AND .CONCLUSIONS 98
REFERENCES 99
CHAPTER 9. Experimentally Comparing Uncertain Inference Systems to Probability 102
1. Abstract 102
2. Introduction 102
3. Outline of Our Method 102
4. Outline of Biases in MYC and TSM 106
5. Experiments on Performance 109
6. Best and Worst Results for MYC and TSM 110
7. Summary 111
8. References 111
PART II: TOOLS 114
CHAPTER 10. KNOWLEDGE ENGINEERING WITHIN A GENERALIZED BAYESIAN FRAMEWORK 116
1. INTRODUCTION 116
2. KNOWLEDGE ENGINEERING WITHIN THE GENERALIZED BAYESIAN FRAMEWORK 117
3. GENERALIZED BAYESIAN INFERENCE 122
4. GENERALIZED BAYESIAN EXPLANATION 125
5. FUTURE WORK 125
6. SUMMARY 126
REFERENCES 126
CHAPTER 11. LEARNING TO PREDICT: AN INDUCTIVE APPROACH 128
1. INTRODUCTION 128
2. BASIC INDUCTIVE COMPONENTS FOR INCREMENTAL LEARNING 131
3. CONCLUSIONS 136
ACKNOWLEDGEMENTS 136
REFERENCES 136
CHAPTER 12. TOWARDS A GENERAL-PURPOSE BELIEF MAINTENANCE SYSTEM 138
1. INTRODUCTION 138
2. DESIGN 138
3. EXAMPLES 142
4. CONCLUSIONS 144
ACKNOWLEDGEMENTS 144
REFERENCES 144
CHAPTER 13. A NON-ITERATIVE MAXIMUM ENTROPY ALGORITHM 146
1 Introduction 146
2 Formal Problem Definition 147
3 The Maximum Entropy Principle 148
4 Iterative Maximum Entropy Methods 149
5 Non-Iterative Techniques 150
6 Acyclic Hypergraphs 151
7 A New Maximum Entropy Method 153
8 Spiegelhalter's Algorithm 157
9 Comparisons 158
10 Conclusions and Open Problems 159
References 159
CHAPTER 14. Propagating Uncertainty in Bayesian Networks by Probabilistic Logic Sampling 162
1. Introduction 162
2. Bayesian Belief Networks 163
3. Dependent evidence and multiply connected networks 165
4. Probabilistic Logic sampling 167
5. Pulse: An Implementation 171
6. Explanation and sensitivity analysis 171
7. Precision and Computational effort 172
8. Improvements to efficiency 174
9. Final remarks 174
Acknowledgments 175
References 175
CHAPTER 15. AN EXPLANATION MECHANISM FOR BAYESIAN INFERENCING SYSTEMS 178
1. INTRODUCTION 178
2. THE GENERALIZED BAYESIAN INFERENCING SYSTEM 179
3. EXPLANATION FACILITIES 180
4. CONCLUDING REMARKS 185
FOOTNOTES AND REFERENCES 186
CHAPTER 16. ON THE RATIONAL SCOPE OF PROBABILISTIC RULE-BASED INFERENCE SYSTEMS 188
1. INTRODUCTION 188
2. BACKGROUND AND NOMENCLATURE 190
3. THE CF LANGUAGE AND ITS RATIONAL INTERPRETATION 191
4. THE CF LANGUAGE AS A SPECIAL CASE OF THE BAYESIAN LANGUAGE 192
5. DISCUSSION 193
6. IMPLICATIONS ON KNOWLEDGE ENGINEERING AND FUTURE RESEARCH 196
7. CONCLUSION 198
8. Appendix: Proofs 198
REFERENCES 201
CHAPTER 17. DAVID: Influence Diagram Processing System for the Macintosh 204
REFERENCES 209
CHAPTER 18. Qualitative Probabilistic Networks for Planning Under Uncertainty 210
1 Introduction 210
2 Probabilistic Networks 211
3 Qualitative Influences 211
4 An Example: The Generic Test/Treat Decision 215
5 Conclusions 219
References 220
CHAPTER 19. ON IMPLEMENTING USUAL VALUES 222
1. INTRODUCTION 222
2. ON POSSIBILITY-PROBABILITY GRANULES 223
3. ON USUAL VALUES AND THEIR REPRESENTATION 224
4. TRANSLATION OF COMPOUND STATEMENTS 225
5. LOGICAL TRANSLATION RULES 226
6. REASONING WITH USUAL VALUES 227
6. ARITHMETIC OPERATIONS WITH USUAL OPERATIONS 228
7. CONCLUSION 229
REFERENCES 229
PART Ill: THEORY 232
CHAPTER 20. SOME EXTENSIONS OF PROBABILISTIC LOGIC 234
1. INTRODUCTION 234
2. EVIDENTIAL LOGIC 235
3. SEMANTICS AS RANDOM VARIABLES 236
4. PROBABILISTIC LOGIC AS CONSISTENT LABELING 238
REFERENCES 239
CHAPTER 21. Belief as Summarization and Meta-Support 242
1. Introduction 242
2. The Network Model 243
3. Computing Belief and Reliability values 245
4. A Network of Cognitive Units 247
5. Conclusions 248
References 248
CHAPTER 22. NON-MONOTONICITY IN PROBABILISTIC REASONING 250
1 Introduction 250
2 Probabilistic Logic 250
3 Non-Monotonic Probabilistic Theories 251
4 Default Inheritance of Probabilities 254
5 Specificity-Prioritized Maximization of Conditional Independence 255
6 Non-Monotonicity in "Evidential" Reasoning 255
7 Graphoids, Influence Diagrams, and Irrelevance 256
8 A Circumscriptive Formalization of (SP) MCI 257
9 Maximum Entropy 258
10 Discussion 259
11 Conclusion 260
12 Directions for Future Research 260
Acknowledgements 261
Notes 261
References 261
CHAPTER 23. A SEMANTIC APPROACH TO NON-MONOTONICENTAILMBNTS 264
1. OVERVIEW 264
2. TRum 265
3. ENTAILMENT 267
4. PROBABILITY 271
5. CONCLUSION 275
ACKNOWLEDGEMENTS 275
REFERENCES 275
CHAPTER 24. KNOWLEDGE 276
1. BACKGROUND. 276
2. SUBJECTIVE MEASURES. 276
3. PROBABILITY 278
4. UNCERTAIN KNOWLEDGE 280
5. DECISION 284
BIBLIOGRAPHY 285
CHAPTER 25. Computing Reference Classes 286
1. Reference Classes. 286
2. Kyburg's Strategy and Its Capabilities. 288
3. Lessons from Implementation 294
4. Concluding Discussion. 299
Acknowledgements 302
References 302
CHAPTER 26. DISTRIBUTED REVISION OF BELIEF COMMITMENT IN COMPOSITE EXPLANATIONS 304
ABSTRACT 304
1. INTRODUCTION 304
2. REVIEW OF BELIEF UPDATING IN BAYESIAN BELIEF NETWORKS 306
3. BELIEF REVISION IN SINGLY-CONNECTED NETWORKS 309
4. COPING WITH LOOPS 316
5. A MEDICAL DIAGNOSIS EXAMPLE 319
CONCLUSIONS 325
ACKNOWLEDGMENT 326
REFERENCES 326
CHAPTER 27. A BACKWARDS VIEW FOR ASSESSMENT 330
1. INTRODUCTION 330
2. INFLUENCE DIAGRAMS 331
3. DETERMINISTIC MODELS 332
4. PROBABILISTIC MODELS 333
5. CONCLUSIONS 335
ACKNOWLEDGEMENTS 336
NOTES 336
REFERENCES 336
CHAPTER 28. PROPAGATION OF BELIEF FUNCTIONS: A DISTRIBUTED APPROACH 338
I. Abstract and Introduction 338
II. Belief Functions 339
II.. Qualitative Markov Trees 340
IV. Propagating Belief Functions in Qualitative Markov Trees 342
V. Conclusion 347
VI. Acknowledgements 347
VII. References 348
CHAPTER 29. GENERALIZING FUZZY LOGIC PROBABILISTIC INFERENCES 350
1. INTRODUCTION 350
2. GENERATING FUNCTIONS FOR BOOLEAN FORMULAS 353
3. COMPUTING PROJECTIONS 364
4. COMPOSITION OF FACES 371
5. A TRANSFORNATION OF ANY BOOLEAN FORMULA TO A FACET OF BN 2 372
6. CONCLUSIONS 373
REFERENCES 373
PART IV: APPLICATIONS 376
CHAPTER 30. THE SUM-AND-LATTICE-POINTS METHOD BASED ON AN EVIDENTIAL-REASONING SYSTEM APPLIED TO THE REAL-TIME VEHICLE GUIDANCE PROBLEM 378
1. INTRODUCTION 378
2. PROBLEM 378
3. RELEVANCE 378
4. THE NECESSITY OF INTRODUCING GENERAL EVIDENTIAL REASONING 379
5. THE GENERAL EVIDENTIAL REASONING MODEL 380
6. APPROACH: SUM-AND-LATTICE-POINTS METHOD 381
7. THE EQUIVALENCE PROOF AND THE TRANSITION GRAPH 382
8. PARALLEL IMPLEMENTATION 382
9. ADVANTAGES 382
10. CONCLUSION 383
REFERENCES 383
CHAPTER 31. Probabilistic Reasoning About Ship Images 384
1. INTRODUCTION 384
2. REASONING ABOUT SHIP IMAGES 385
3. A SIMPLE PROTOTYPE 387
4. SCALING UP TO MORE REALISTIC PROBLEMS 388
5. THE BMS APPROACH 389
6. CONCLUSIONS 391
Acknowledgements 391
REFERENCES 392
CHAPTER 32. Information and Multi-Sensor Coordination 394
1 Introduction 394
2 A Team-Theoretic Formulation of Multi-Sensor Systems 395
3 Simulation Studies 400
4 Evaluation and Speculation 403
5 Conclusions and Future Research 405
References 406
CHAPTER 33. Planning, Scheduling, and Uncertainty in the Sequence of Future Events 408
Abstract 408
1. Statement of the Problem 408
2. Candidate Solutions 409
3. The Proposed Solution 410
4. Empirical Results 411
5. Conclusion 412
References 412
CHAPTER 34. EVIDENTIAL REASONING IN A COMPUTER VISION SYSTEM 416
ABSTRACT 416
1. INTRODUCTION 416
2. A SET-THEORETICAL EVIDENTIAL REASONING APPROACH TO COMPUTERVISION 417
3. PROGRAMMING RESULTS 421
4. CONCLUSION 425
ACKNOWLEDGEMENTS 425
REFERENCES 425
CHAPTER 35. BAYES IAN INFERENCE FOR RADAR IMAGERY BASED SURVEILLANCE 426
I. INTRODUCTION 426
2. EVIDENTIAL ACCRUAL 429
3. APPROXIMATE CONFLICT RESOLUTION 431
4. ISSUES 433
ACKNOWLEDGEMENTS 433
REFERENCES 433
CHAPTER 36. A CAUSAL BAYESIAN MODEL FOR THE DIAGNOSIS OF APPENDICITIS 436
1. INTRODUCTION AND OVERVIEW 436
2. STANDARD BAYESIAN ASSUMPTIONS AND CRITIQUES 437
3. THE CAUSAL BAYESIAN MODEL 438
4. THE KNOWLEDGE ENGINEERING 440
5. IMPLEMENTING THE MODEL 441
6. TESTING THE MODEL 443
7. WORK IN PROGRESS 444
8. CONCLUSIONS 445
REFERENCES 445
CHAPTER 37. Estimating Uncertain Spatial Relationships in Robotics 448
1 Introduction 448
2 The Stochastic Map 449
3 Reading the Map 454
4 Building the Map 460
5 Developed Example 466
6 Discussion and Conclusions 469
Appendix A 470
Relationships Using Euler Angles 470
Relationships Using Roll, Pitch and Yaw Angles 472
References 473

Produkt-Details

• Erscheint lt. Verlag: 28.6.2014
• Sprache: englisch
• Themenwelt: Informatik ► Theorie / Studium ► Künstliche Intelligenz / Robotik
• Themenwelt: Technik
• ISBN-10: 1-4832-9653-9 / 1483296539
• ISBN-13: 978-1-4832-9653-1 / 9781483296531