Heckerman / Mamdani Uncertainty in Artificial Intelligence

Proceedings of the Ninth Conference on Uncertainty in Artificial Intelligence, The Catholic University of America, Washington, D.C. 1993
1. Auflage 2014
ISBN: 978-1-4832-1451-1
Verlag: Elsevier Science & Techn.
Format: PDF
Kopierschutz: 1 - PDF Watermark

E-Book, Englisch, 552 Seiten, Web PDF

Uncertainty in Artificial Intelligence contains the proceedings of the Ninth Conference on Uncertainty in Artificial Intelligence held at the Catholic University of America in Washington, DC, on July 9-11, 1993. The papers focus on methods of reasoning and decision making under uncertainty as applied to problems in artificial intelligence (AI) and cover topics ranging from knowledge acquisition and automated model construction to learning, planning, temporal reasoning, and machine vision. Comprised of 66 chapters, this book begins with a discussion on causality in Bayesian belief networks before turning to a decision theoretic account of conditional ought statements that rectifies glaring deficiencies in classical deontic logic and forms a sound basis for qualitative decision theory. Subsequent chapters explore trade-offs in constructing and evaluating temporal influence diagrams; normative engineering risk management systems; additive belief-network models; and sensitivity analysis for probability assessments in Bayesian networks. Automated model construction and learning as well as algorithms for inference and decision making are also considered. This monograph will be of interest to both students and practitioners in the fields of AI and computer science.

Autoren/Hrsg.

Heckerman, David

Mamdani, Abe

Inhaltsverzeichnis

1;Front Cover;1
2;Uncertainty in
Artificial
Intelligence;4
3;Copyright Page;5
4;Table of Contents;6
5;Preface;10
6;Acknowledgements;11
7;Part 1: Foundations;12
7.1;Chapter 1. Causality in Bayesian Belief Networks;14
7.1.1;Abstract;14
7.1.2;1 INTRODUCTION;14
7.1.3;2 SIMULTANEOUS EQUATIONS MODELS;16
7.1.4;3 CAUSALITY IN BAYESIAN BELIEF NETWORKS;19
7.1.5;4 CONCLUSION;21
7.1.6;Acknowledgments;22
7.1.7;References;22
7.2;Chapter 2. From Conditional Oughts to Qualitative Decision Theory;23
7.2.1;Abstract;23
7.2.2;1 INTRODUCTION;23
7.2.3;2 INFINITESIMAL
PROBABILITIES, RANKING
FUNCTIONS, CAUSAL
NETWORKS, AND ACTIONS;24
7.2.4;3 SUMMARY OF RESULTS;25
7.2.5;4 FROM UTILITIES AND BELIEFS TO GOALS AND ACTIONS;25
7.2.6;5 COMBINING ACTIONS AND OBSERVATIONS;27
7.2.7;6 RELATIONS TO OTHER ACCOUNTS;29
7.2.8;7 CONCLUSION;31
7.2.9;Acknowledgements;31
7.2.10;References;31
8;Part 2: Applications and Empirical Comparisons;32
8.1;Chapter 3. A Probabilistic Algorithm for Calculating Structure:Borrowing from Simulated Annealing;34
8.1.1;Abstract;34
8.1.2;1 MOLECULAR STRUCTURE;34
8.1.3;2 THE DATA REPRESENTATION;35
8.1.4;3 EXPERIMENTS PERFORMED;37
8.1.5;4 RESULTS;38
8.1.6;5 DISCUSSION;39
8.1.7;6 CONCLUSIONS;41
8.1.8;Acknowledgements;42
8.1.9;References;42
8.2;Chapter 4. A Study of Scaling Issues in Bayesian Belief Networks for Ship Classification;43
8.2.1;Abstract;43
8.2.2;1 Introduction;43
8.2.3;2 Overview;43
8.2.4;3 Network Structure;45
8.2.5;4 Integration of Belief Values;46
8.2.6;5 Discussion;47
8.2.7;6 Conclusion;48
8.2.8;References;48
8.3;CHAPTER 5. TRADEOFFS IN CONSTRUCTING AND EVALUATING TEMPORAL INFLUENCE DIAGRAMS;51
8.3.1;Abstract;51
8.3.2;1 INTRODUCTION;51
8.3.3;2 TEMPORAL BAYESIAN NETWORKS;52
8.3.4;3 TID CONSTRUCTION FROM KNOWLEDGE BASES;53
8.3.5;4 DOMAIN-SPECIFIC TIME-SERIES MODELS;54
8.3.6;5 MODEL SELECTION APPROACHES;55
8.3.7;6 EVALUATING TRADEOFFS;57
8.3.8;7 RELATED LITERATURE;57
8.3.9;8 CONCLUSIONS;58
8.3.10;Acknowledgements;58
8.3.11;References;58
8.4;Chapter 6. End-User Construction of Influence Diagrams for Bayesian Statistics;59
8.4.1;Abstract;59
8.4.2;1 INTRODUCTION;59
8.4.3;2 STATISTICAL MODEL;60
8.4.4;3 SEMANTIC INTERFACE: THE PATIENT-FLOW DIAGRAM;61
8.4.5;4 METADATA-STATE DIAGRAM: THE COHORT-STATE DIAGRAM;62
8.4.6;5 CONSTRUCTION STEPS;62
8.4.7;6 IMPLEMENTATION;63
8.4.8;7 OTHER WORK;64
8.4.9;8 CONCLUSION;64
8.4.10;Acknowledgments;65
8.4.11;References;65
8.5;Chapter 7. On Considering Uncertainty and Alternatives in Low-Level Vision;66
8.5.1;Abstract;66
8.5.2;1 INTRODUCTION;66
8.5.3;2 REGIONS, SEGMENTS, AND SEGMENTATIONS;68
8.5.4;3 SEGMENT-LEVEL UNCERTAINTY;68
8.5.5;4 SEGMENTATION-LEVEL UNCERTAINTY;69
8.5.6;5 REGION-LEVEL UNCERTAINTY;71
8.5.7;6 OBTAINING PRIORS;72
8.5.8;7 ALGORITHMS;72
8.5.9;8 AN EXPERIMENTAL EXAMPLE;73
8.5.10;9 CONCLUSION;73
8.5.11;Acknowledgement;73
8.5.12;References;73
8.6;Chapter 8. Forecasting Sleep Apnea with Dynamic Network Models;75
8.6.1;Abstract;75
8.6.2;1 INTRODUCTION;75
8.6.3;2 RELATED WORK;76
8.6.4;3 THE DYNAMIC NETWORK MODEL;76
8.6.5;4 THE DYNEMO IMPLEMENTATION;77
8.6.6;5 THE SLEEP-APNEA FORECASTING PROBLEM;79
8.6.7;6 CONCLUSIONS;81
8.6.8;Acknowledgments;81
8.6.9;References;81
8.7;Chapter 9. Normative Engineering Risk Management Systems;83
8.7.1;Abstract;83
8.7.2;1 INTRODUCTION;83
8.7.3;2 ENGINEERING RISK MANAGEMENT SYSTEMS;83
8.7.4;3 ADVANCED RISK MANAGEMENT SYSTEM PROJECT;84
8.7.5;4 NORMATIVE SYSTEM OVERVIEW;85
8.7.6;5 NORMATIVE SYSTEM ACTIVITIES;86
8.7.7;6 RESEARCH ISSUES;89
8.7.8;7 CONCLUSIONS;89
8.7.9;Acknowledgements;89
8.7.10;References;89
8.8;Chapter 10. Diagnosis of Multiple Faults: A Sensitivity Analysis;91
8.8.1;Abstract;91
8.8.2;1 INTRODUCTION;91
8.8.3;2 THE MODELS;92
8.8.4;3 EXPERIMENTAL DESIGN;93
8.8.5;4 RESULTS AND DISCUSSION;94
8.8.6;5 Acknowledgment;94
8.8.7;References;94
9;Part 3: Knowledge Acquisition, Modelling, and Explanation;100
9.1;Chapter 11. Additive Belief-Network Models;102
9.1.1;Abstract;102
9.1.2;1 INTRODUCTION;102
9.1.3;2 ADDITIVE MODEL;102
9.1.4;3 SIGNIFICANCE OF ADDITIVEDECOMPOSITION;103
9.1.5;4 FITTING ADDITIVEBELIEF-NETWORK MODELS;105
9.1.6;5 INFERENCE ALGORITHM;107
9.1.7;6 IMPLEMENTATION RESULTS;108
9.1.8;7 CONCLUSIONS;108
9.1.9;Acknowledgments;108
9.1.10;References;108
9.2;Chapter 12. Parameter adjustment in Bayes networks.The generalized noisy OR-gate;110
9.2.1;Abstract;110
9.2.2;1 INTRODUCTION;110
9.2.3;2 PARAMETER ADJUSTMENT;111
9.2.4;3 THE GENERALIZED NOISYOR-GATE;113
9.2.5;4 CONCLUSIONS;116
9.2.6;Acknowledgments;116
9.2.7;References;116
9.3;Chapter 13. A fuzzy relation-based extension of Reggiavs relational model fordiagnosis handling uncertain and incomplete information;117
9.3.1;Abstract;117
9.3.2;1 INTRODUCTION;117
9.3.3;2 RELATIONAL APPROACH : THECOMPLETELY INFORMED CASE;117
9.3.4;3 THE CASE OF INCOMPLETEINFORMATION;118
9.3.5;4 LINK WITH REGGIA ET AL.'SAPPROACH;119
9.3.6;5 GRADED UNCERTAINTY VS.GRADED INTENSITY OFPRESENCE;120
9.3.7;6 A NEW MODEL BASED ONTWOFOLD FUZZY SETS;121
9.3.8;7 CONCLUDING REMARKS;123
9.3.9;References;124
9.4;CHapter 14. Dialectic reasoning with inconsistent information;125
9.4.1;Abstract;125
9.4.2;1 INTRODUCTION;125
9.4.3;2 ARGUMENTATION;126
9.4.4;3 CONSTRUCTING ARGUMENTS;128
9.4.5;4 ACCEPTABILITY CLASSES;128
9.4.6;5 LINGUISTIC QUALIFIERS;129
9.4.7;6 USING PRIORITIES;130
9.4.8;7 Final remarks;131
9.4.9;Acknowledgement;131
9.4.10;References;131
9.5;Chapter 15. Causal Independence for Knowledge Acquisition and Inference;133
9.5.1;Abstract;133
9.5.2;1 INTRODUCTION;133
9.5.3;2 A TEMPORAL DEFINITION OFCAUSAL INDEPENDENCE;134
9.5.4;3 A BELIEF-NETWORKREPRESENTATION OF CAUSALINDEPENDENCE;134
9.5.5;4 A REAL-WORLD EXAMPLE;135
9.5.6;5 IMPROVEMENTS IN THEREPRESENTATION FORINFERENCE;136
9.5.7;6 A LIMITATION OF THEREPRESENTATION FORINFERENCE;137
9.5.8;7 AN OBSERVATION ABOUTGENERALITY;137
9.5.9;References;137
9.6;Chapter 16. Utility-Based Abstraction and Categorization;139
9.6.1;Abstract;139
9.6.2;1 INTRODUCTION;139
9.6.3;2 ACTIONS UNDERUNCERTAINTY;139
9.6.4;3 ABSTRACTION BYUTILITY-BASED SIMILARITY;140
9.6.5;4 POLYNOMIAL COMPUTATIONOF ABSTRACTIONS;141
9.6.6;5 EXAMPLES OF UTILITY-BASED
ABSTRACTION;142
9.6.7;6 DECISIONS WITHABSTRACTIONS;144
9.6.8;7 SUMMARY AND CONCLUSIONS;145
9.6.9;References;146
9.7;Chapter 17. Sensitivity Analysis forProbability Assessments in Bayesian Networks;147
9.7.1;Abstract;147
9.7.2;1 INTRODUCTION;147
9.7.3;2 COMPUTING SENSITIVITY VALUES;148
9.7.4;3. EXAMPLE;150
9.7.5;4. INCORPORATING DIRECTESTIMATES OF TARGETDISTRIBUTIONS;151
9.7.6;5. DISCUSSION;152
9.7.7;References;152
9.7.8;APPENDIX: PROOFS OF RESULTS;153
9.8;Chapter 18. Causal Modeling;154
9.8.1;Abstract;154
9.8.2;1 INTRODUCTION;154
9.8.3;2 INFORMAL RELATION TO D-GRAPHS;154
9.8.4;3 CAUSAL MODELS;155
9.8.5;4 CONCLUSION;162
9.8.6;5 FURTHER RESEARCH;162
9.8.7;References;162
9.9;Chapter 19. Some Complexity Considerations in the Combination of BeliefNetworks;163
9.9.1;Abstract;163
9.9.2;1 INTRODUCTION;163
9.9.3;2 THE GENERAL APPROACH;163
9.9.4;3 PREVIOUS WORK;164
9.9.5;4 FURTHER THEORETICALDEVELOPMENT;165
9.9.6;5 COMPLEXITY ANALYSIS;167
9.9.7;6 SUMMARY;169
9.9.8;References;169
9.10;Chapter 20. Deriving a Minimal /-map of a Belief Network Relative to aTarget Ordering of its Nodes;170
9.10.1;Abstract;170
9.10.2;1 INTRODUCTION ANDMOTIVATION;170
9.10.3;2 ILLUSTRATIONS;171
9.10.4;3 PRELIMINARIES;172
9.10.5;4 DERIVING DAG'S;173
9.10.6;5 PROOF OF CORRECTNESS;174
9.10.7;6 SUMMARY;176
9.10.8;Acknowledgement;176
9.10.9;References;176
9.11;Chapter 21. Probabilistic Conceptual Network:A Belief Representation Scheme for Utility-Based Categorization;177
9.11.1;1 Introduction;177
9.11.2;2 Integrating Uncertainty and Categorical Knowledge;178
9.11.3;3 An Application in Automated Machining;178
9.11.4;4 Probabilistic Conceptual Network;179
9.11.5;5 Model Construction;182
9.11.6;6 Related Work;182
9.11.7;7 Conclusion;183
9.11.8;Acknowledgements;184
9.11.9;Reference;184
9.12;Chapter 22. Reasoning about the Value of Decision-Model Refinement: Methods and Application;185
9.12.1;Abstract;185
9.12.2;1 Introduction;185
9.12.3;2 Expected Values of Decision-Model Refinement;186
9.12.4;3 Control of Refinement;191
9.12.5;4 Discussion and Related Work;192
9.12.6;5 Summary and Conclusions;193
9.12.7;Reference;193
9.13;Chapter 23. Mixtures of Gaussians and Minimum Relative Entropy Techniques for Modeling Continuous Uncertainties;194
9.13.1;Abstract;194
9.13.2;1 INTRODUCTION;194
9.13.3;2 INFERENCE AND DECISION MAKING
WITH MIXTURES OF GAUSSIANS: AN
INFLUENCE DIAGRAM APPROACH;195
9.13.4;3 TRANSFORMATIONS TOWARD A DESIRED DISTRIBUTION;197
9.13.5;4 FITTING MIXTURE DISTRIBUTIONS WITH THE EM ALGORITHM;198
9.13.6;5 SELECTING THE MIXTURE SIZE;199
9.13.7;6 CONCLUDING REMARKS;200
9.13.8;References;201
9.14;Chapter 24. Valuation Networks and Conditional Independence;202
9.14.1;Abstract;202
9.14.2;1 INTRODUCTION;202
9.14.3;2 VBSs AND CONDITIONAL INDEPENDENCE;203
9.14.4;3 VALUATION NETWORKS;205
9.14.5;4 COMPARISON;207
9.14.6;5 CONCLUSION;209
9.14.7;Acknowledgments;209
9.14.8;References;209
9.15;Chapter 25. Relevant Explanations: Allowing Disjunctive Assignments;211
9.15.1;Abstract;211
9.15.2;1 INTRODUCTION;211
9.15.3;2 GIB EXPLANATION;212
9.15.4;3 GIB-MAP ALGORITHM;215
9.15.5;4 DISCUSSION;216
9.15.6;5 SUMMARY;216
9.15.7;References;218
9.16;Chapter 26. A Generalization of the Noisy-Or Model;219
9.16.1;Abstract;219
9.16.2;1 INTRODUCTION;219
9.16.3;2 THE GENERALIZED MODEL;220
9.16.4;3 CHARACTERIZING P(X\V);221
9.16.5;4 INTERESTING SPECIAL CASES;221
9.16.6;5 COMPUTING P(X\U);222
9.16.7;6 EXAMPLES;223
9.16.8;7 IMPLEMENTATION;226
9.16.9;Acknowledgements;226
9.16.10;References;226
10;Part 4: Automated Model Construction and Learning;228
10.1;Chapter 27. Using First-Order Probability Logic for the Construction of Bayesian Networks;230
10.1.1;Abstract;230
10.1.2;1 Introduction;230
10.1.3;2 Representing General ProbabilisticKnowledge;231
10.1.4;3 Representing Bayesian Networks;231
10.1.5;4 Simple Model Construction;232
10.1.6;5 More General Model Construction;233
10.1.7;6 Conclusions and Future Work;236
10.1.8;References;236
10.2;Chapter 28. Representing and Reasoning With Probabilistic Knowledge: A Bayesian Approach;238
10.2.1;Abstract;238
10.2.2;1 INTRODUCTION;238
10.2.3;2 REPRESENTING PROBABILISTIC KNOWLEDGE;238
10.2.4;3 PROBABILISTIC INFERENCE;242
10.2.5;4 RELATED WORK;243
10.2.6;5 CONCLUSIONS;245
10.2.7;References;245
10.3;Chapter 29. Graph-Grammar Assistance for Automated Generation of Influence Diagrams;246
10.3.1;Abstract;246
10.3.2;1 MODELING OF DECISIONS;246
10.3.3;2 GRAPH GRAMMARS;246
10.3.4;3 A GRAPH GRAMMAR FOR MEDICAL DECISIONS;248
10.3.5;4 DISCUSSION;249
10.3.6;Acknowledgments;251
10.3.7;References;251
10.4;Chapter 30.Using Causal Information and Local Measures to Learn Bayesian Networks;254
10.4.1;Abstract;254
10.4.2;1 Introduction;254
10.4.3;2 Learning Bayesian Networks;255
10.4.4;3 Localization of the Description Length;257
10.4.5;4 Incorporating Partial Domain Knowledge;258
10.4.6;5 Searching for the Best Constrained Network;258
10.4.7;6 Experiments;259
10.4.8;7 Refinement of Existent Networks;261
10.4.9;References;261
10.5;Chapter 31. Minimal Assumption Distribution Propagation in Belief Networks;262
10.5.1;Abstract;262
10.5.2;1 INTRODUCTION;262
10.5.3;2 PREVIOUS WORK;263
10.5.4;3 FRAMEWORK FOR LEARNING QBNS;263
10.5.5;4 PROPAGATING DISTRIBUTIONS;264
10.5.6;5 EXAMPLE;267
10.5.7;6 CONCLUSION;269
10.5.8;Acknowledgements;269
10.5.9;References;269
10.6;Chapter 32. An Algorithm for the Construction of Bayesian Network Structures from Data;270
10.6.1;Abstract;270
10.6.2;1 INTRODUCTION;270
10.6.3;2 MOTIVATION;271
10.6.4;3 DISCUSSION OF THE ALGORITHM;272
10.6.5;4 PRELIMINARY RESULTS;273
10.6.6;5 SUMMARY AND OPEN PROBLEMS;275
10.6.7;Acknowledgements;275
10.6.8;References;276
10.7;Chapter 33. A Construction of Bayesian Networks from Databases Based on an MDL Principle;277
10.7.1;Abstract;277
10.7.2;1 INTRODUCTION;277
10.7.3;2 DISCUSSION WITHOUTASSUMING BAYESIAN BELIEFNETWORKS;278
10.7.4;3 DISCUSSION ASSUMINGBAYESIAN BELIEF NETWORKS;280
10.7.5;4 Concluding Remarks;282
10.7.6;Acknowledgements;282
10.7.7;References;282
10.7.8;Appendix A: Proof of Theorem 2;283
10.7.9;Appendix .: Proof of Theorem 8;284
10.8;Chapter 34. Knowledge-Based Decision Model Construction for Hierarchical Diagnosis: A Preliminary Report;285
10.8.1;Abstract;285
10.8.2;1 Introduction;285
10.8.3;2 Skeleton of our Methodology;286
10.8.4;3 The Functional Component;286
10.8.5;4 The Bridge fault component;288
10.8.6;5 The Meta-Level Component;288
10.8.7;6 An Example;290
10.8.8;7 Related research;290
10.8.9;8 Conclusion;292
10.8.10;Acknowledgments;292
10.8.11;References;292
11;Part 5: Algorithms for Inference and Decision Making;294
11.1;Chapter 35. A Synthesis of Logical and Probabilistic Reasoning for Program Understanding and Debugging;296
11.1.1;Abstract;296
11.1.2;1 INTRODUCTION;296
11.1.3;2 PROBLEM FOCUS;296
11.1.4;3 OPERATING SYSTEMS AND PROGRAM UNDERSTANDING;297
11.1.5;4 LOGICAL ANALYSIS OF EXECUTION PATHS;298
11.1.6;5 UNCERTAINTY ABOUT ERRORS ON PATHS;299
11.1.7;6 UNCERTAINTY AND PATH IDENTIFICATION;301
11.1.8;7 SUMMARY;302
11.1.9;References;302
11.2;Chapter 36. An Implementation of a Method
for Computing the Uncertainty
in Inferred Probabilities in Belief Networks;303
11.2.1;Abstract;303
11.2.2;1 INTRODUCTION;303
11.2.3;2 STATISTICAL VARIANCE AND DIRICHLET DISTRIBUTIONS;304
11.2.4;3 THE APPROXIMATE PROPAGATIONMETHOD;304
11.2.5;4 THE MONTE CARLO INTEGRATIONMETHOD;305
11.2.6;5 COMPARISON OF METHODS;305
11.2.7;6 DISCUSSION AND CONCLUSION;306
11.2.8;References;307
11.3;Chapter 37.
Incremental Probabilistic Inference;312
11.3.1;Abstract;312
11.3.2;1 Introduction;312
11.3.3;2 Desiderata;312
11.3.4;3 Term Computation - a new task definition;313
11.3.5;4 Error Estimates;316
11.3.6;5 Making Term Computation Incremental;317
11.3.7;6 Experimental Evaluation;317
11.3.8;7 Discussion;318
11.3.9;8 Conclusion;318
11.3.10;References;318
11.4;Chapter 38. Deliberation Scheduling for Time-Critical Sequential Decision Making;320
11.4.1;Abstract;320
11.4.2;1 Introduction;320
11.4.3;2 Deliberation Scheduling;321
11.4.4;3 Precursor Deliberation;322
11.4.5;4 Recurrent Deliberation;324
11.4.6;5 Related Work and Conclusions;327
11.4.7;Acknowledgements;327
11.4.8;References;327
11.5;Chapter 39. Intercausal Reasoning with Uninstantiated Ancestor Nodes;328
11.5.1;Abstract;328
11.5.2;1 INTRODUCTION;328
11.5.3;2 QUALITATIVE PROBABILISTIC NETWORKS;329
11.5.4;3 UNINSTANTIATED ANCESTOR NODES;329
11.5.5;4 PRODUCT SYNERGY II;330
11.5.6;5 NOISY-OR DISTRIBUTIONS;332
11.5.7;6 CONCLUSION;333
11.5.8;APPENDIX: PROOFS;333
11.5.9;Acknowledgments;336
11.5.10;References;336
11.6;Chapter 40. Inference Algorithms for Similarity Networks;337
11.6.1;Abstract;337
11.6.2;1 INTRODUCTION;337
11.6.3;2 DEFINITION OF SIMILARITY NETWORKS;337
11.6.4;3 TWO TYPES OF SIMILARITY NETWORKS;339
11.6.5;4 INFERENCE USING SIMILARITY NETWORKS;340
11.6.6;5 INFERENTIAL AND DIAGNOSTIC COMPLETENESS;341
11.6.7;6 OPEN PROBLEMS;345
11.6.8;Acknowledgments;345
11.6.9;References;345
11.7;Chapter 41. Two Procedures for Compiling Influence Diagrams;346
11.7.1;ABSTRACT;346
11.7.2;1.0 INTRODUCTION;346
11.7.3;2.0 COMPILING DECISION NETWORKS;347
11.7.4;3.0 APPLICATIONS;351
11.7.5;4.0 FUTURE WORK;351
11.7.6;Acknowledgement;351
11.7.7;References;351
11.8;Chapter 42. An efficient approach for finding the MPE in belief networks;353
11.8.1;Abstract;353
11.8.2;1 Introduction;353
11.8.3;2 The algorithm for finding the MPE;354
11.8.4;3 Finding the / MPEs in belief networks;355
11.8.5;4 The MPE for a subset of variables in belief networks;358
11.8.6;5 Related work;359
11.8.7;6 Conclusion;360
11.8.8;References;360
11.9;Chapter 43. A Method for Planning Given Uncertain and Incomplete Information;361
11.9.1;Abstract;361
11.9.2;1 Introduction;361
11.9.3;2 State Representation;363
11.9.4;3 Reduction Operator;364
11.9.5;4 Representation of Plans;365
11.9.6;5 The Basic Planning Algorithm;366
11.9.7;6 Plan Reapplication;366
11.9.8;7 Super-Plans;367
11.9.9;8 Results;367
11.9.10;9 Conclusion;368
11.9.11;Acknowledgments;369
11.9.12;References;369
11.10;Chapter 44. heuse of conflicts in searching Bayesian networks;370
11.10.1;Abstract;370
11.10.2;1 Introduction;370
11.10.3;2 Background;371
11.10.4;3 Searching possible worlds;371
11.10.5;4 Estimating the Probabilities;372
11.10.6;5 Discussion;373
11.10.7;6 A Diagnosis Example;373
11.10.8;7 Search Strategy and Conflicts;375
11.10.9;8 Comparison with other systems;377
11.10.10;9 Conclusion;377
11.10.11;Acknowledgements;377
11.10.12;References;378
11.11;Chapter 45. GALGO: A Genetic ALGOrithm Decision Support Tool for ComplexUncertain Systems Modeled with Bayesian Belief Networks;379
11.11.1;Abstract;379
11.11.2;1 INTRODUCTION;379
11.11.3;2 BELIEF NETWORKS;379
11.11.4;3 GENETIC ALGORITHMS;380
11.11.5;4 THE ALGORITHM BEHIND GALGO;380
11.11.6;5 EXAMPLES;382
11.11.7;6 EXPERIMENTAL RESULTS;383
11.11.8;7 DISCUSSION;384
11.11.9;8. REFERENCES;386
11.12;Chapter 46. Using Tree-Decomposable Structures to ApproximateBelief Networks;387
11.12.1;Abstract;387
11.12.2;1 INTRODUCTION;387
11.12.3;2 BELIEF NETWORKS AND BELIEFTREES;387
11.12.4;3 OPTIMAL TREE-DECOMPOSABLENETWORKS;390
11.12.5;4 CONCLUSIONS AND FUTURERESEARCH;393
11.12.6;Acknowledgements;393
11.12.7;References;393
11.13;Chapter 47. Using Potential Influence Diagrams forProbabilistic Inference and Decision Making;394
11.13.1;Abstract;394
11.13.2;1 INTRODUCTION;394
11.13.3;2 CONDITIONAL INFLUENCEDIAGRAMS;394
11.13.4;3 POTENTIAL INFLUENCEDIAGRAMS;396
11.13.5;4 COMPOUND OPERATIONS ANDALGORITHMS;398
11.13.6;5 CONCLUSIONS AND FUTURERESEARCH;400
11.13.7;Acknowledgments;401
11.13.8;References;401
11.14;Chapter 48. Deciding Morality of Graphs is ...-complete;402
11.14.1;Abstract;402
11.14.2;1 INTRODUCTION;402
11.14.3;2 PRELIMINARIES;403
11.14.4;3 SUFFICIENT CONDITIONS FORMORALITY;403
11.14.5;4 NECESSARY CONDITIONS FORMORALITY;403
11.14.6;5 ANOTHER SUFFICIENTCONDITION;404
11.14.7;6 COMPLEXITY ANALYSIS;404
11.14.8;References;407
11.15;Chapter 49. Incremental computation of the value of perfect information instepwise-decomposable influence diagrams;411
11.15.1;Abstract;411
11.15.2;1 INTRODUCTION;411
11.15.3;2 STEPWISE-DECOMPOSABLEINFLUENCE DIAGRAMS;412
11.15.4;3 CONDENSING SDID'S;413
11.15.5;4 COMPUTING THE VALUE OFPERFECT INFORMATION;416
11.15.6;5 CONCLUSIONS;418
11.15.7;A cknowledgement;418
11.15.8;References;418
12;Part 6: Qualitative Reasoning;420
12.1;Chapter 50. Argumentative inference in uncertainand inconsistent knowledge bases;422
12.1.1;Abstract;422
12.1.2;1. Introduction;422
12.1.3;2. Arguments in Flat Knowledge bases;422
12.1.4;3. Properties of »—;424
12.1.5;4. Arguments in prioritized knowledgebases;426
12.1.6;5. Paraconsistent-Like Reasoning inLayered Knowledge Bases;428
12.1.7;6. Combining knowledge bases;429
12.1.8;7. Conclusion;430
12.1.9;References;430
12.2;Chapter 51. Argument Calculus and Networks;431
12.2.1;Abstract;431
12.2.2;1 INTRODUCTION;431
12.2.3;2 ARGUMENT DATABASES;431
12.2.4;3 INDEPENDENCE;433
12.2.5;4 ARGUMENT NETWORKS;434
12.2.6;5 COMPUTING ARGUMENTS;434
12.2.7;6 APPLICATIONS OFARGUMENT NETWORKS;435
12.2.8;CONCLUSION;438
12.2.9;ACKNOWLEDGEMENT;438
12.2.10;References;438
12.3;Chapter 52. Argumentation as a General Framework for Uncertain Reasoning;439
12.3.1;Abstract;439
12.3.2;1 INTRODUCTION;439
12.3.3;2 ARGUMENTATION;439
12.3.4;3 ARGUMENTATION CONSEQUENCERELATION;440
12.3.5;4 FLATTENING AND AGGREGATION;440
12.3.6;5 THE ARGUMENTATION THEOREMPROVER;440
12.3.7;6 AGGREGATION CRITERIA;442
12.3.8;7 QUANTITATIVE UNCERTAINTYCALCULI;442
12.3.9;8 ARGUMENTATION ANDDEFEASIBILITY;443
12.3.10;9 META-ARGUMENTATION;443
12.3.11;10 RESOLVING CONFLICTS ANDMAKING DECISIONS;443
12.3.12;11 CONCLUSION;444
12.3.13;Acknowledgements;444
12.3.14;References;444
12.4;Chapter 53. On reasoning in networks with qualitative uncertainty;446
12.4.1;Abstract;446
12.4.2;1 INTRODUCTION;446
12.4.3;2 A N EW QUALITATIVEAPPROACH;446
12.4.4;3 QUALITATIVE CHANGES INSIMPLE NETWORKS;448
12.4.5;4 A COMPARISON OF THETHREE FORMALISMS;449
12.4.6;5 QUALITATIVE CHANGES INMORE COMPLEX NETWORKS;449
12.4.7;6 INTEGRATION THROUGH QUALITATIVE CHANGE;451
12.4.8;7 DISCUSSION;452
12.4.9;8 SUMMARY;453
12.4.10;A cknowledgement s;453
12.4.11;References;453
12.5;Chapter 54. Qualitative Measures of Ambiguity;454
12.5.1;Abstract;454
12.5.2;1 Introduction;454
12.5.3;2 An Ambiguity Measure;455
12.5.4;3 Relationship between Ambiguityand Other Measures of Uncertainty;456
12.5.5;4 Conclusion;458
12.5.6;Appendix;458
12.5.7;References;461
13;Part 7: Interpretation and Comparisonof Approaches forReasoning Under Uncertainty;462
13.1;Chapter 55. A BAYESIAN VARIANT OF SHAFER'S COMMONALITIES FOR MODELLING UNFORESEEN EVENTS;464
13.1.1;Abstract;464
13.1.2;1 INTRODUCTION;465
13.1.3;2 MAPPING EVENTS INTO THE POWER SET OF F;466
13.1.4;3 EXPECTED UTILITY WITH NORMALIZEDCOMMONALITIES;467
13.1.5;4 CONCLUSIONS;469
13.1.6;Acknowledgements;470
13.1.7;References;470
13.2;Chapter 56. The Probability of a Possibility:Adding Uncertainty to Default Rules;472
13.2.1;Abstract;472
13.2.2;1 Introduction;472
13.2.3;2 Belief Revision and Possibilistic Logic;474
13.2.4;3 Counterfactual Probabilities;475
13.2.5;4 Generalized Imaging9;477
13.2.6;5 Concluding Remarks;478
13.2.7;Acknowledgements;479
13.2.8;References;479
13.3;Chapter 57. Possibilistic decreasing persistence;480
13.3.1;Abstract;480
13.3.2;1 Introduction;480
13.3.3;2 Background on possibilistic logic;481
13.3.4;3 Possibilistic decreasing persistence:the extrapolation problem;482
13.3.5;4 From qualitative to quantitativeaxioms for persistence schemata;484
13.3.6;5 Inferring nonmonotonic conclusionsfrom decreasing persistence;486
13.3.7;6 Concluding remarks;486
13.3.8;Acknowledgement;487
13.3.9;References;487
13.4;Chapter 58. DISCOUNTING AND COMBINATION OPERATIONS INEVIDENTIAL REASONING;488
13.4.1;Abstract;488
13.4.2;1 INTRODUCTION;488
13.4.3;2 HOW TO DISCOUNT THEEVIDENTIAL FUNCTIONS;488
13.4.4;3 DISCOUNTING A N DCOMBINATION OPERATIONS;490
13.4.5;4 SUMMARY;495
13.4.6;References;495
13.5;Chapter 59. Probabilistic Assumption-Based Reasoning;496
13.5.1;Abstract;496
13.5.2;1 MODELING PROPOSITIONALSYSTEMS;496
13.5.3;2 A LOGICAL V I EW OFSUPPORTS OF HYPOTHESES;497
13.5.4;3 COMPUTINGCHARACTERISTIC CLAUSES;499
13.5.5;4 COMPUTING DEGREES OFSUPPORT;500
13.5.6;5 CONCLUSION;502
13.5.7;References;502
13.6;Chapter 60. Partially Specified Belief Functions;503
13.6.1;Abstract;503
13.6.2;1 INTRODUCTION;503
13.6.3;2 BASIC CONCEPTS ANDRESULTS;504
13.6.4;3 THE MINIMUM SPECIFICITYAND LEAST COMMITMENTPRINCIPLES;506
13.6.5;4 THE FOCUSING PRINCIPLE;508
13.6.6;5 CONCLUSIONS;510
13.6.7;Acknowledgements;510
13.6.8;References;510
13.7;Chapter 61. Jeffrey's rule of conditioning generalized to belief functions.;511
13.7.1;Abstract:;511
13.7.2;1. Jeffrey's rule in probability theory.;511
13.7.3;2. Revision versus focusing;512
13.7.4;3. Jeffrey's rule applied to belieffunctions;514
13.7.5;4. Conclusions.;515
13.7.6;References.;516
13.8;Chapter 62. Inference with Possibilistic Evidence;517
13.8.1;Abstract;517
13.8.2;1 INTRODUCTION;517
13.8.3;2 POSSIBILISTIC EVIDENCE;518
13.8.4;3 INFERENCE WITHPOSSIBILISTIC EVIDENCE;520
13.8.5;4 SUMMARY AND DISCUSSION;524
13.8.6;References;525
13.9;Chapter 63. Constructing Lower Probabilities;526
13.9.1;Abstract;526
13.9.2;1 Introduction;526
13.9.3;2 An Alternative Construal of fi;526
13.9.4;3 Some Special Cases;527
13.9.5;4 Acknowledgements;529
13.9.6;References;529
13.10;Chapter 64. Belief Revision in Probability Theory;530
13.10.1;Abstract;530
13.10.2;1 INTRODUCTION;530
13.10.3;2 PROPAGATION VS. REVISION;530
13.10.4;3 EXPLICIT CONDITION VS. IMPLICITCONDITION;531
13.10.5;4 UPDATING VS. REVISION;533
13.10.6;5 A DEFECT OF THE BAYESIANAPPROACH;533
13.10.7;6 AN EXAMPLE;534
13.10.8;7 SUMMARY;536
13.10.9;Acknowledgements;537
13.10.10;References;537
13.11;Chapter 65. The Assumptions Behind Dempster's Rule;538
13.11.1;Abstract;538
13.11.2;1 INTRODUCTION;538
13.11.3;2 SOURCE STRUCTURES ANDBELIEF FUNCTIONS;538
13.11.4;3 COMBINATION RULES;539
13.11.5;4 DEMPSTER'S RULE OFCOMBINATION;540
13.11.6;5 BAYESIAN CONDITIONING;542
13.11.7;6 CONSTRAINTS ANDASSUMPTIONS ON C-RULES;542
13.11.8;7 DISCUSSION;543
13.11.9;Acknowledgements;544
13.11.10;References;544
13.12;Chapter 66. A Belief-Function Based Decision Support System;546
13.12.1;Abstract;546
13.12.2;1. INTRODUCTION;546
13.12.3;2. THEORETICAL BACKGROUND- TRANSFERABLE BELIEF MODEL;547
13.12.4;3. A BELIEF FUNCTION BASEDDECISION SUPPORT SYSTEM;547
13.12.5;4. AN EXAMPLE;548
13.12.6;3. CONCLUSIONS;552
13.12.7;Acknowledgments;552
13.12.8;References;552
14;Author Index;554