Uncertainty Management in Expert Systems
IEEE Expert: Intelligent Systems and Their Applications
| Hi-index | 0.00 |
MUM IS A KNOWLEDGE-BASED CONSULTATION SYSTEM DESIGNED TO MANAGER THE UNCERTAINTY INHERENT IN MEDICAL DIAGNOSIS (THE ACRONYM STANDS FOR MANAGE- MENT OF UNCERTAINTY IN MEDICINE). MANAGING UNCERTAINTY MEANS PLANNING ACTIONS TO MINIMIZE UNCERTAINTY OR ITS CONSEQUENCES. THUS IT IS A CONTROL PROBLEM--AN ISSUE FOR THE COMPONENT OF A KNOWLEDGE SYSTEM THAT DECIDES HOW TO PROCEED FROM AN UNCERTAIN STATE OF A PROBLEM. UNCERTAINTY CAN BE MANAGED BY MANY STRATEGIES, DEPENDING ON THE KIND OF PROBLEM ONE IS TRYING TO SOLVE. THESE INCLUDE ASKING FOR EVIDENCE, HEDGING ONE''S BETS, DECIDING ARBITRARILY AND BACKTRACKING ON FAILURE, DIVERSIFICATION OR RISK-SHARING, AND WORST-CASE ANALYSIS. THE FACILITY WITH WHICH A CONSULTATION SYSTEM SUCH AS MUM MANAGES UNCERTAINTY IS EVIDENT IN THE QUESTIONS IT ASKS: IT SHOULD ASK ALL NECESSARY QUESTIONS, NO UNNECESSARY QUESTIONS, AND IT SHOULD ASK ITS QUESTIONS IN THE RIGHT ORDER. THESE CONDITIONS, ESPECIALLY THE LAST ONE, PRECLUDE UNIFORM AND INFLEXIBLE CONTROL STRATEGIES. THEY PROMPTED THE DEVELOPMENT OF THE MUM ARCHITECTURE IN WHICH CONTROL DECISIONS ARE TAKEN BY REASONING ABOUT FEATURES OF EVIDENCE AND SOURCES OF UNCERTAIN- TY.