Single step mapping in topological order of the queued class and instance frames of a semantic network to a static working memory
First Claim
1. A method of operating a processing means to map updated information upon a call to access from a dynamic frame-based semantic network having class frames and instance frames, with each of said instance frames being associated with at least one of said class frames, to a static working memory having a plurality of working memory elements, each said working memory element corresponding to one of said class frames and instance frames, the method comprising the steps of:
- (a) utilizing the processing means to create one class queue of the class frames and one instance queue of the instance frames in a memory means to obtain queued class frames and instance frames having the updated information to be mapped;
(b) utilizing the processing means to mark each of the queued class frames of the class queue to order the queued class frames of the class queue to reflect a currently proper sequence of updating the working memory elements;
said step (b) further includes setting a counter corresponding to each of the queued class frames, each said counter having a value based on a topological sort of the queued class frames; and
(c) utilizing the processing means to update the working memory elements of the working memory corresponding to the queued class frames and queued instance frames with the updated information to be mapped based on the order of the queued class frames of the class queue to obtain an updated working memory, said working memory elements corresponding to one of said queued class frames and said queued instance frames associated with said one of said queued class frames being updated in a single step.
1 Assignment
0 Petitions
Accused Products
Abstract
Method and apparatus are provided for mapping information from a dynamic frame-based semantic network to a static working memory by utilizing a topological sorting algorithm which processes at least one queue of frames. Preferably, the algorithm is utilized with filtering to avoid unnecessary updates. The algorithm makes a marking pass and an updating pass across the queued frames. The marking pass orders a queue of class frames. The updating pass updates working memory elements of the working memory corresponding to the ordered class frames and working memory elements which correspond to instances of the class. After the updating pass, the working memory elements reflect local and inheritable slots and values from the semantic network. Such mapping allows the working memory elements to be further processed such as by matching against rules.
40 Citations
12 Claims
-
1. A method of operating a processing means to map updated information upon a call to access from a dynamic frame-based semantic network having class frames and instance frames, with each of said instance frames being associated with at least one of said class frames, to a static working memory having a plurality of working memory elements, each said working memory element corresponding to one of said class frames and instance frames, the method comprising the steps of:
-
(a) utilizing the processing means to create one class queue of the class frames and one instance queue of the instance frames in a memory means to obtain queued class frames and instance frames having the updated information to be mapped; (b) utilizing the processing means to mark each of the queued class frames of the class queue to order the queued class frames of the class queue to reflect a currently proper sequence of updating the working memory elements;
said step (b) further includes setting a counter corresponding to each of the queued class frames, each said counter having a value based on a topological sort of the queued class frames; and(c) utilizing the processing means to update the working memory elements of the working memory corresponding to the queued class frames and queued instance frames with the updated information to be mapped based on the order of the queued class frames of the class queue to obtain an updated working memory, said working memory elements corresponding to one of said queued class frames and said queued instance frames associated with said one of said queued class frames being updated in a single step. - View Dependent Claims (2, 3, 4, 5, 6)
-
-
7. In a data processing system including processing means and memory means including a static working memory having a plurality of working memory elements operatively coupled to the processing means, apparatus for mapping updated information upon a call to access from a dynamic frame-based semantic network having class frames and instance frames, with each of said instance frames being associated with at least one of said class frames, to the static working memory having a plurality of working memory elements, each said working memory element corresponding to one of said class frames and instance frames, the apparatus comprising:
-
(a) means for creating one class queue of the class frames and one instance queue of the instance frames in the memory means to obtain queued class frames and instance frames having the updated information to be mapped; (b) means for marking each of the queued class frames of the class queue to order the queued class frames of the class queue to reflect a currently proper sequence of updating the working memory elements;
said means (b) further includes means for setting a counter corresponding to each of the queued class frames, each said counter having a value based on a topological sort of the queued class frames; and(c) means for updating the working memory elements of the working memory corresponding to the queued class frames and queued instance frames with the updated information to be mapped based on the order of the queued class frames of the class queue to obtain an updated working memory, said working memory elements corresponding to one of said queued class frames and said queued instance frames associated with said one of said queued class frames being updated in a single step. - View Dependent Claims (8, 9, 10, 11, 12)
-
Specification