Applications of circumscription to formalizing common-sense knowledge
Artificial Intelligence
A study of aboutness in information retrieval
Artificial Intelligence Review
A Study of Belief Revision in the Context of Adaptive Information Filtering
ICSC '99 Proceedings of the 5th International Computer Science Conference on Internet Applications
Intelligent Text Handling Using Default Logic
ICTAI '96 Proceedings of the 8th International Conference on Tools with Artificial Intelligence
IJCAI'85 Proceedings of the 9th international joint conference on Artificial intelligence - Volume 1
Hi-index | 0.00 |
When we traverse over the Internet to search information, we sometimes change an object which we are looking for, in other word, change a retrieval goal according to additional information obtained during search. In this paper, we give a formalization of such a goal change by prioritized abduction and extend the method to a “retrieval command adviser”. We assume a chain of logical rules to satisfy a purpose of the retrieval where bottom conditions are retrieval goals represented as abducible propositions. Abducing these retrieval goals leads to accomplishment of the retrieval purpose. Moreover, we introduce another kind of abducible propositions to represent applicability of logical rules. This applicability abducible is attached to a logical rule and used to detect which rules are used. Then, together with priority over these applicability abducibles to express strength of the associated rules, we determine which rules should preferably be applicable and hence we infer the most appropriate retrieval command which is derived from the most preferable rules. If the observed information is changed, then the applicability of rules is changed and according to the applicability abducibles, we change our retrieval goal. We believe that this mechanism explains a retrieval goal change. Then, we extend the method to construct a “retrieval command adviser” which suggests better retrieval command than the initial command given by a user. We translate the above logical rules into another form which is used for predicting a purpose of give a command. When a user give a command then the system predict the purpose of a command using the newly introduced rules and then according to the initial logic rules and priorities, we infer the most appropriate retrieval command.