Uncertainty in Artificial Intelligence (eBook)

Proceedings of the Seventh Conference on Uncertainty in Artificial Intelligence, UCLA, at Los Angeles, July 13-15, 1991

Piero Patrone Bonissone, Bruce D'Ambrosio, Philippe Smets (Herausgeber)

eBook Download: PDF

2014 | 1. Auflage
445 Seiten
Elsevier Science (Verlag)
978-1-4832-9856-6 (ISBN)

Infos

Uncertainty Proceedings 1991

Inhalt

Front Cover 1
Uncertainty in Artificial Intelligence 4
Copyright Page 5
Table of Contents 8
Preface 6
Chapter 1. ARCO1: An Application of Belief Networks to the Oil Market 12
Abstract 12
1 Introduction 12
2 Domain Specifics 12
3 Model Variables 13
4 Scenarios 14
5 Forecasts 15
6 Conclusions 16
7 Acknowledgements 17
8 References 17
Chapter 2. "Conditional Inter-Causally Independent" node distributions, a property of "noisy-or" models 20
Abstract 20
1 EVIDENCE NODES THAT ARE COMMON TO MULTIPLE PARENTS 20
2 CONSTRUCTIVE SOLUTION OF THE BINARY VARIABLE INTER-CAUSAL DEPENDENCY 25
3 DISCUSSION 27
Acknowledgements 27
References 27
Chapter 3. Combining Multiple-valued Logics in Modular Expert Systems 28
Abstract 28
1 INTRODUCTION 28
2 ENTAILMENT SYSTEMS 29
3 A CLASS OF MULTIPLE-VALUED LOGICS FOR THE UNCERTAINTY MANAGEMENT IN RULE-BASED EXPERT SYSTEMS 30
4 INFERENCE PRESERVING MAPS BETWEEN MV-LOGICS 31
5 CONCLUSIONS AND FUTURE WORK 35
Acknowledgements 36
References 36
Chapter 4. Constraint Propagation with Imprecise Conditional Probabilities 37
Abstract 37
1 INTRODUCTION 37
2 STATEMENT OF THE PROBLEM 38
3 A LINEAR PROGRAMMING METHOD 39
4 GENERALIZED BAYES' THEOREM 39
5 LOCAL INFERENCE RULES 40
6 A CONSTRAINT PROPAGATION BASED ON INFERENCE RULES 41
7 AN EXAMPLE 42
8 CONJUNCTION AND DISJUNCTION 42
9 INDEPENDENCE ASSUMPTIONS 43
10 CONCLUSION 44
Acknowledgements 45
References 45
Chapter 5. BAYESIAN NETWORKS APPLIED TO THERAPY MONITORING 46
Abstract 46
1. INTRODUCTION 46
2. HIGH-LEVEL VIEW OF THE MODEL 47
3. INFERENCE 48
4. COMPUTING THE INFERENCES VIA STOCHASTIC SIMULATION 49
5. SPECIFIC MODEL FOR CYTOTOXIC CHEMOTHERAPY MONITORING IN BREAST CANCER 49
7. CONCLUSIONS 52
Acknowledgements 53
References 53
Chapter 6. Some Properties of Plausible Reasoning 55
Abstract 55
1 INTRODUCTION 55
2 NOTATION 56
3 THEORY 57
4 EXAMPLES 59
5 CONCLUSION 60
References 61
Chapter 7. Theory Refinement on Bayesian Networks 63
Abstract 63
1 Introduction 63
2 Bayesian Networks 64
3 Partial Bayesian networks 65
4 Representing alternative Bayesian networks 66
5 Theory Refinement 67
6 Extensions 69
7 Conclusion 70
Acknowledgements 70
References 70
Chapter 8. COMBINATION OF UPPER AND LOWER PROBABILITIES 72
Abstract 72
1 INTRODUCTION 72
2 'A PRIORI' INFORMATION 73
3 EVIDENTIAL INFORMATION 74
4 COMBINATION OF 'A PRIORI AND EVIDENTIAL INFORMATION 76
Acknowledgments 79
References 79
Chapter 9. A Probabilistic Analysis of Marker-Passing Techniques for Plan-Recognition 80
Abstract 80
1 Introduction 80
2 Probabilistic Schema Evaluation 81
3 Probabilistic Schema Selection 81
4 Path Calculations 84
5 Results 86
Acknowledgements 87
References 87
Chapter 10. Symbolic Probabilistic Inference with Continuous Variables 88
Abstract 88
1 Introduction 88
2 Overview of the SPI Algorithm 89
3 The SPI with Continuous Variables Algorithm 90
4 Conclusion 92
References 92
Chapter 11. Symbolic Probabilistic Inference with Evidence Potential 93
Abstract 93
1 Introduction 93
2 Evidence Potential Algorithm 94
3 Symbolic Inference with Evidence Potential 94
4 Examples 95
5 Conclusion 96
References 96
Chapter 12. A Bayesian Method for Constructing Bayesian Belief Networks from Databases 97
Abstract 97
1 INTRODUCTION 97
2 METHODS 98
3 PRELIMINARY RESULTS 103
4 SUMMARY OF THE LEARNING METHOD AND RELATED WORK 103
Acknowledgements 104
References 104
Chapter 13. Local Expression Languages for Probabilistic Dependence: a preliminary report 106
Abstract 106
1 Introduction 106
2 Overview of SPI 106
3 Local Expression Languages for Probabilistic Knowledge 108
4 Discussion 112
5 Conclusion 112
Acknowledgements 113
References 113
Chapter 14. Symbolic Decision Theory and Autonomous Systems 114
Abstract 114
1 INTRODUCTION 114
2 SYMBOLIC DECISION MAKING UNDER UNCERTAINTY 115
3 AUTONOMOUS DECISION MAKING UNDER UNCERTAINTY 118
Acknowledgements 121
References 121
Chapter 15. A REASON MAINTENANCE SYSTEM DEALING WITH VAGUE DATA 122
Abstract 122
INTRODUCTION 122
MANY-VALUED LOGICS AND RESOLUTION 122
DEFINITION OF A FUZZY TRUTH MAINTENANCE SYSTEM 124
CONCLUSION 127
Acknowledgements 127
References 127
Chapter 16. Advances in Probabilistic Reasoning 129
Abstract 129
1 Introduction 129
2 Representation and Inference 129
3 Knowledge Acquisition/Representation 133
4 Generalized Similarity Networks 135
5 Summary 136
References 137
Chapter 17. Probability Estimation in face of Irrelevant Information 138
Abstract 138
1 INTRODUCTION 138
2 THE UNDERLYING MODEL 139
3 THE ESTIMATION PROBLEM 140
4 JUSTIFICATION AND EXTENSIONS 142
5 COMPARISON TO OTHER WORK 143
6 CONCLUSION 144
Acknowledgments 144
References 144
Chapter 18. An Approximate Nonmyopic Computation for Value of Information 146
Abstract 146
1 INTRODUCTION 146
2 VALUE-OF-INFORMATION COMPUTATIONS FOR DIAGNOSIS 146
3 MYOPIC ANALYSIS 147
4 NONMYOPIC ANALYSIS 149
5 VALUE OF INFORMATION FOR A SUBSET OF EVIDENCE 149
6 RELAXATION OF THE ASSUMPTIONS 150
7 SUMMARY AND CONCLUSIONS 152
Acknowledgments 152
References 152
Chapter 19. Search-based Methods to Bound Diagnostic Probabilities in Very Large Belief Nets 153
Abstract 153
1 INTRODUCTION 153
2 QMR AND INTERNIST 154
3 QMR-BN: A PROBABILISTIC INTERPRETATION OF QMR 154
4 INFERENCE ALGORITHMS 155
5 NOTATION 156
6 RELATIVE PROBABILITY AND MARGINAL EXPLANATORY POWER 156
7 NEGATIVE PRODUCT SYNERGY AND THE MEP THEOREM 156
8 BOUNDS ON THE PROBABILITY OF EXTENSIONS 157
9 SEARCH METHOD 158
10 OBTAINING ABSOLUTE PROBABILITIES 158
11 PERFORMANCE OF TOPN 159
CONCLUSIONS 160
Acknowledgements 160
References 160
Chapter 20. Chapter Time-Dependent Utility and Action Under Uncertainty 162
Abstract 162
1 INTRODUCTION 162
2 A LIMITED REASONER 162
3 TIME-DEPENDENT UTILITY 164
4 PROTOS IN ACTION 166
5 SUMMARY 168
Acknowledgments 169
References 169
Chapter 21. Non-monotonic Reasoning and the Reversibility of Belief Change 170
Abstract 170
1 INTRODUCTION 170
2 BELIEF CHANGE AND INFERENCE 170
3 SEMANTICS FOR BELIEF CHANGE 171
4 ITERATED BELIEF CHANGE AND REVERSIBILITY 172
5 DISCUSSION 174
Acknowledgements 174
References 174
Chapter 22. Belief and Surprise - A Belief-Function Formulation 176
Abstract 176
1 INTRODUCTION 176
2 BELIEF FUNCTIONS AS A GENERAL FORMALIZATION MECHANISM 178
3 A CASE STUDY 181
4 DISCUSSION 182
5 CONCLUSION 183
Acknowledgements 183
Appendix - logical formulas and subsets of 183
References 184
Chapter 23. Evidential Reasoning in a Categorial Perspective: Conjunction and Disjunction of Belief Functions 185
Abstract 185
0 INTRODUCTION 185
1 FROM THE DYNAMICS OF BELIEFS TO CATEGORIES OR ... VICE VERSA 186
2 CATEGORIES OF "BELIEFS" 187
3 DISJUNCTIONS AND CONJUNCTIONS 189
4 COPRODUCTS AND CONJUNCTIONS 189
5 PRODUCTS AND DISJUNCTIONS 190
6 SEPARABLE BELIEF FUNCTIONS 191
7 CONCLUSIONS 191
Acknowledgments 192
References 192
Chapter 24. Reasoning with Mass Distributions 193
Abstract 193
1 INTRODUCTION 193
2 REPRESENTING KNOWLEDGE WITH MASS DISTRIBUTIONS 193
3 THE CONCEPT OF SPECIALIZATION 195
4 SPECIALIZATION MATRICES 196
5 CONCLUSIONS 198
Chapter 25. A Logic of Graded Possibility and Certainty Coping with Partial Inconsistency 199
ABSTRACT 199
1 INTRODUCTION 199
2 POSSIBILISTIC LOGIC : LANGUAGE AND SEMANTICS 200
3 AUTOMATED DEDUCTION IN POSSIBILISTIC LOGIC 203
CONCLUSION 206
Acknowledgements 206
References 206
Chapter 26. Conflict and Surprise: Heuristics for Model Revision 208
Abstract 208
1 INTRODUCTION 208
2 BACKGROUND 208
3 THEORETICAL FRAMEWORK 210
4 REBUTTALS 212
5 RARE CASES 214
6 DISCUSSION 214
Acknowledgements 215
References 215
Chapter 27. Reasoning under Uncertainty: Some Monte Carlo Results 216
Abstract 216
1 INTRODUCTION 216
2 METHOD 216
3 RESULTS 217
4 DISCUSSION 221
References 222
Chapter 28. Representation Requirements for Supporting Decision Model Formulation 223
Abstract 223
1 Introduction 223
2 An Example 224
3 The Decision Making Process 224
4 Summary of Inference Patterns and Representation Requirements 226
5 A Representation Design 227
6 Supporting General Inferences 228
7 Related Work 229
8 Discussion and Conclusion 229
Acknowledgments 230
References 230
Chapter 29. A Language for Planning with Statistics 231
Abstract 231
1 INTRODUCTION 231
2 KNOWLEDGE REPRESENTATION 232
3 INFERENCE 233
4 PLANNING 235
5 CONCLUSION 237
Acknowledgments 238
References 238
Chapter 30. A Modification to Evidential Probability 239
Abstract 239
1 Overview of the Problem 239
2 The Proposed Solution 240
3 Conclusions 242
Acknowledgments 242
References 242
Chapter 31. Investigation of Variances in Belief Networks 243
Abstract 243
1 INTRODUCTION 243
2 PRELIMINARY ASSUMPTIONS 245
3 DETERMINING THE VARIANCES IN INFERRED PROBABILITIES 246
4 OBTAINING AN UPPERBOUND FOR THE PRIOR VARIANCES 249
5 FUTURE RESEARCH 252
References 252
Chapter 32. A Sensitivity Analysis of Pathfinder: A Follow-up Study 253
Abstract 253
1 INTRODUCTION 253
2 DETAILS OF THE ANALYSIS 254
3 THE INITIAL STUDY 254
4 THE FOLLOW-UP STUDY 255
5 CONCLUSIONS 257
Acknowledgments 259
References 259
Chapter 33. Non-monotonic Negation in Probabilistic Deductive Databases 260
Abstract 260
1 Introduction 260
2 Syntax and Uses of General Probabilistic Logic Programs 261
3 Background: Fixpoint Theory for Pf-programs 262
4 Stability of Formula Functions 263
5 Stable Classes of Formula Functions 264
6 Discussion 265
7 Conclusions 266
Acknowledgements 266
References 266
Chapter 34. Management of Uncertainty in the Multi-Level Monitoring and Diagnosis of the Time of Flight Scintillation Array 268
Abstract 268
1 INTRODUCTION 268
2 BACKGROUND LITERATURE 269
3 TIME OF FLIGHT SCINTILLATION ARRAY 269
4 SYSTEM ARCHITECTURE 269
5 MANAGEMENT OFUNCERTAINTY AT THE MONITORING LEVEL 270
6 MANAGEMENT OF UNCERTAINTY AT THE STRUCTURAL REASONING LEVEL 271
7 MANAGEMENT OF UNCERTAINTY AT THE BEHAVIORAL REASONING LEVEL 271
8 IMPLEMENTATION 272
9 SUMMARY 272
Acknowledgements 273
References 273
Chapter 35. Integrating Probabilistic Rules into Neural Networks: A Stochastic EM Learning Algorithm 275
Abstract 275
1 INTRODUCTION 275
2 PROBABILISTIC NETWORKS 276
3 MAXIMUM LIKELIHOOD ESTIMATION 277
4 THE STOCHASTIC EM-ALGORITHM 278
5 DISCUSSION 280
Acknowledgements 280
References 280
Chapter 36. Representing Bayesian Networks within Probabilistic Horn Abduction 282
Abstract 282
1 Introduction 282
2 Probabilistic Horn Abduction 282
3 Representing Bayesian networks 284
4 Best-first abduction 286
5 Causation 287
6 Comparison with Other Systems 287
7 Conclusion 287
Acknowledgements 289
References 289
Chapter 37. DYNAMIC NETWORK UPDATING TECHNIQUES FOR DIAGNOSTIC REASONING 290
Abstract 290
1 INTRODUCTION 290
2 DYNAMICS OF DIAGNOSTIC REASONING UNDER UNCERTAINTY 291
3 SYSTEM ARCHITECTURE 291
4 MODEL CONSTRUCTION HEURISTICS 292
5 MODEL UPDATING 294
6 CONCLUSIONS 297
References 297
Chapter 38. High Level Path Planning with Uncertainty 298
Abstract 298
1 INTRODUCTION 298
2 U–GRAPH 299
3 PATH PLANNING WITH UNCERTAINTY 299
4 A FORMAL DEFINITION OF PATH PLANNING 300
5 RELATED WORK 304
6 CONCLUSION AND FUTURE WORK 305
Acknowledgements 305
References 305
Chapter 39. Formal Model of Uncertainty for Possibilistic Rules 306
OVERVIEW 306
1 POSSIBILITY DISTRIBUTIONS AND MEASURES 306
2 INFORMATION FUNCTIONS IN POSSIBILITY THEORY 307
3 DESIGN OF CONTINUOUS POSSIBILITY INFORMATION 308
4 PROPERTIES OF CONTINUOUS INFORMATION MEASURES 308
5 PRINCIPLE OF MAXIMUM UNCERTAINTY 309
REFERENCES 309
Chapter 40. Deliberation and its Role in the Formation of Intentions* 311
Abstract 311
1 INTRODUCTION 311
2 OVERVIEW 312
3 POSSIBLE WORLDS MODEL 312
4 DECISION TREES AND GOAL WORLDS 315
5 DELIBERATION AND INTENTIONS 316
6 CONCLUSIONS 317
References 318
Chapter 41. Handling Uncertainty during Plan Recognitionin Task-Oriented Consultation Systems 319
Abstract 319
1 INTRODUCTION 319
2 THE INFERENCE MECHANISM 320
3 THE PROBABILITY OF AN INTERPRETATION OF THE DISCOURSE 321
4 STRENGTH OF INFERENCES 323
5 INFORMATION CONTENT AND ITS USE 324
6 EXAMPLES 325
7 CONCLUSIONS 326
Acknowledgments 326
References 326
Chapter 42. TRUTH AS UTILITY: A CONCEPTUAL SYNTHESIS 327
Abstract 327
1 Introduction 327
2 Possible Worlds and Desirabilities 328
3 Desirability and Preference 329
4 Combination of Preference Functions 331
5 Possibility and Necessity 331
6 Preference, Similarity, and Fuzzy Logic 332
AckNowledgements 333
References 333
Chapter 43. PULCINELLAA General Tool for Propagating Uncertainty in Valuation Networks 334
Abstract 334
1. INTRODUCTION 334
2. THEORETICAL BACKGROUND 335
3. PULCINELLA 336
4. EXAMPLES 338
5. DISCUSSION 340
6. CONCLUSIONS 341
Acknowledgements 342
References 342
Chapter 44. Structuring Bodies of Evidence 343
Abstract 343
1 INTRODUCTION 343
2 BASIC NOTIONS IN EVIDENCE THEORY 343
3 PROPOSAL OF STRUCTURES 344
4 DEMPSTER RULE OF COMBINATION 347
5 LOCAL PROPAGATION OF INFORMATION 348
6 CONCLUSION 349
Acknowledgements 349
References 349
Chapter 45. On the Generation of Alternative Explanations with Implications for Belief Revision 350
Abstract 350
1 INTRODUCTION 350
2 CONSTRAINT SYSTEMS 351
3 GENERATING ALTERNATIVE EXPLANATIONS 352
4 BAYESIAN NETWORKS 355
5 DISCUSSION 357
Acknowledgments 358
References 358
Chapter 46. Completing Knowledge by Competing Hierarchies 359
Abstract 359
1 Introduction 359
2 The knowledge base 359
3· The control strategy 360
4. The application to a multi-hierarchical knowledge base 362
5. Discussion 363
Acknowledgements 363
References 363
Chapter 47. A Graph-Based Inference Method for Conditional Independence 364
Abstract 364
1. INTRODUCTION 364
2. NOTATION AND BASIC CONCEPTS 364
3. MULTIPLE UNDIRECTED GRAPHS 365
4. GRAPHICAL REPRESENTATION OF THE GRAPHOID AXIOMS 366
5. EXTENSIONS TO THE GRAPHICAL OPERATIONS 367
6. EXAMPLES 367
7. CONCLUSIONS 370
Acknowledgements 370
References 370
Chapter 48. A Fusion Algorithm for Solving Bayesian Decision Problems 372
Abstract 372
1 INTRODUCTION 372
2 A DIABETES DIAGNOSIS PROBLEM 372
3 VALUATION-BASED SYSTEM REPRESENTATION 373
4 SOLVING A VBS 376
5 A FUSION ALGORITHM 377
6 CONCLUSIONS 378
Acknowledgements 380
References 380
Chapter 49. Algorithms for Irrelevance-Based Partial MAPs 381
Abstract 381
1 INTRODUCTION 381
2 IB-MAP ALGORITHM 384
3 d-IB MAP ALGORITHM 387
4 FUTURE WORK 387
5 SUMMARY 388
Acknowledgements 388
References 388
Chapter 50. About Updating 389
Abstract 389
1. CONDITIONING RULES FOR BELIEF FUNCTIONS 389
2. THE SCENARIO: THE VOTING INTENTIONS STUDY 392
3. CONDITIONING 392
4. BELIEFS INDUCED BY THE PROPORTIONS 395
5. CONCLUSIONS 396
Acknowledgements 396
Bibliography 396
Chapter 51. Compressed Constraints in Probabilistic Logic and Their Revision 397
Abstract 397
1. PROLIFERATION OF WORLDS 397
2. OVERVIEW AND EXAMPLE 398
3. COMPRESSION USING KNOWLEDGE AND USING SEARCH 398
4. EXPRESSING THE CONSTRAINTS 400
5. REVISION WITH CONDITIONALS 400
6. AN EXAMPLE OF REVISION 401
7. REVISION USING POSTERIORS 402
8· CONCLUSIONS 402
Literature Cited 402
Chapter 52. Detecting Causal Relations in the Presence of Unmeasured Variables 403
Abstract 403
1 Introduction 403
2 Results 403
3 Can Theorem 3 Be Strengthened? 405
4 Appendix 406
Acknowledgements 407
References 408
Chapter 53. A Method for Integrating Utility Analysis into an Expert System for Design Evaluation under Uncertainty 409
Abstract 409
1. INTRODUCTION 409
2. INTEGRATION OF USER-DEFINED EVALUATION FUNCTION INTO EXPERT SYSTEM 410
3. EXAMPLE: AUTOMOTIVE BUMPER MATERIAL SELECTION KBS 413
4. CONCLUSIONS 415
Acknowledgment 415
References 415
Chapter 54. From Relational Databases to Belief Networks 417
Abstract 417
1 INTRODUCTION 417
2 RELATIONAL DATABASES 417
3 BELIEF NETWORKS 419
4 INITIAL DISTRIBUTIONS 422
5 CONCLUSIONS 423
Acknowledgement 424
References 424
Chapter 55. A Monte-Carlo Algorithm for Dempster-Shafer Belief 425
Abstract 425
1 INTRODUCTION 425
2 THE MONTE-CARLO ALGORITHM 425
3 COMPUTATION TIME 426
4 EXPERIMENTAL RESULTS 426
5 THE GENERALISED ALGORITHM 427
6 DISCUSSION 427
Acknowledgements 428
References 428
Chapter 56. Compatibility of Quantitative and Qualitative Representations of Belief 429
Abstract 429
1 INTRODUCTION 429
2 QUANTITATIVE BELIEF MEASURES 430
3 PREFERENCE RELATIONS VERSUS QUANTITATIVE BELIEF MEASURES 431
4 CONCLUSION 434
Acknowledgements 435
References 435
Chapter 57. An Efficient Implementation of Belief Function Propagation 436
Abstract 436
1 INTRODUCTION 436
2 SOME BASIC CONCEPTS ABOUT BELIEF FUNCTION NETWORKS 436
3 BELIEF FUNCTION PROPAGATION USING LOCAL COMPUTATION 437
4 A More Efficient Implementation 439
5 UPDATING MESSAGES 441
6 CONCLUSIONS 443
Acknowledgements 443
References 443
Chapter 58. A Non-Numeric Approach to Multi-Criteria/Multi-Expert Aggregation Based on Approximate Reasoning 444
Abstract 444
1. Introduction 444
2. PROBLEM FORMULATION 445
3. A Non-Numeric Technique Multi-Criteria Aggregation 445
4. Combining Expert's Opinions 446
5. CONCLUSION 448
6. REFERENCES 448
Chapter 59. Why Do We Need Foundations for Modelling Uncertainties? 449
1 What Are Foundations? 449
2 Do We Need Foundations At All? 449
3 Testability 450
4 Proliferation and Communication 450
5 Considering Foundations 450
6 What Are We Trying to Do? 451
7 What Are We Talking About? 451
8 Little but the Truth 451
9 More of the Truth 452
10 Usefulness 452
11 Practise and Theory 452
12 A Garden 452
Acknowledgments 453
References 453
Author Index 454

Produkt-Details

• Erscheint lt. Verlag: 28.6.2014
• Sprache: englisch
• Themenwelt: Informatik ► Theorie / Studium ► Künstliche Intelligenz / Robotik
• Themenwelt: Naturwissenschaften
• ISBN-10: 1-4832-9856-6 / 1483298566
• ISBN-13: 978-1-4832-9856-6 / 9781483298566