Methods for the construction and maintenance of a knowledge representation system

• US 20040220969A1
• Filed: 02/02/2004
• Published: 11/04/2004
• Est. Priority Date: 06/08/2000
• Status: Active Grant

• Current Assignee: Ingenuity Systems, Inc. (Qiagen NV)
• Original Assignee: Ingenuity Systems, Inc. (Qiagen NV)
• Inventors: Richards, Daniel R., Cho, Raymond J., Felciano, Ramon M., Chen, Richard O.
• Granted Patent: US 7,577,683 B2
• Time in Patent Office: Days
• US Class Current: 707/104.1
• CPC Class Codes: G06F 16/367 Ontology
• CPC Class Codes: G06N 5/022 Knowledge engineering; Knowledge acquisition
• CPC Class Codes: Y10S 707/99945 Object-oriented database structure processing
• CPC Class Codes: Y10S 707/99948 Application of database or data structure, e.g. distributed, multimedia, or image

Abstract

Methods for constructing and maintaining knowledge representation systems are disclosed herein. The knowledge representation system is initially organized and populated using knowledge engineers. After the initial organization, scientific domain experts digest and structure source texts for direct entry into the knowledge representation system using templates created by the knowledge engineers. These templates constrain both the form and content of the digested information, allowing it to be entered directly into the knowledge representation system. Although knowledge engineers are available to evaluate and dispose of those instances when the digested information cannot be entered in the form required by the templates, their role is much reduced from conventional knowledge representation system construction methods. The methods disclosed herein permit the construction and maintenance of a much larger knowledge representation system than could be constructed and maintained using known methods.

Claims

1. A method for acquiring facts to populate a knowledge representation system (KRS), the method comprising the steps of:

defining and organizing the KRS as an ontology;

creating at least one template, the template accepting a predetermined type of fact, wherein the template formats the predetermined type of fact to be consistent with the organization of the ontology;

acquiring at least one structured fact from a predetermined knowledge source by selecting and completing the template with information found in the predetermined knowledge source; and

inserting the at least one structured fact into the KRS to create a knowledge base.

2. The method of claim 1, further comprising the step of verifying that the at least one completed template conforms to both the organization and permitted content of the ontology.

3. The method of claim 1, wherein the KRS is a frame-based KRS.

4. The method of claim 1, further comprising the step of modifying existing templates and creating additional new templates if facts found in the knowledge sources cannot be accommodated by existing templates.

5. The method of claim 4 wherein the ontology is modified to accept new fact types.

6. The method of claim 1 wherein content scientists knowledgeable in the field of knowledge represented by the ontology complete the at least one template.

7. The method of claim 4 wherein both content scientists and knowledge engineers cooperate to modify the fact templates and the ontology.

8. The method of claim 1 wherein the ontology comprises varying levels of abstraction of biological concepts and the at least one template includes one or more user interface objects corresponding to one or more of the varying levels of abstraction of biological concepts.

9. The method of claim 8 wherein the at least one template includes a plurality of user interface objects corresponding to a respective plurality of abstraction levels of the ontology wherein the plurality of abstraction levels comprises a structured representation of a biological fact.

10. The method of claim 1 wherein the at least one template is completed by a user interface that constrains user data entry based upon a set of valid entries.

11. The method of claim 10, wherein the user interface includes pull-down menus which constrain the information that can be extracted from the predetermined knowledge source and placed in the template.

12. The method of claim 8 wherein the ontology is organized to store genomics information.

13. A frame-based Knowledge Representation System (KRS) populated with facts, the facts having been entered into the system by the step of:

transferring a plurality of structured facts from completed fact templates KRS to form a knowledge base, the structured facts being derived from natural language information sources;

wherein the KRS is an ontology having varying levels of abstraction of biological concepts and the structured facts correspond to one or more of the varying levels of abstraction.

14. The KRS of claim 13 wherein the natural language information sources comprise at least one of text-based and symbolic biological information sources.

15. The KRS of claim 13 wherein the fact templates structure and constrain a fact extracted from the information sources according to a user interface that constrains user data entry based upon a set of valid entries consistent with the structure and content of the ontology.

16. The KRS of claim 15 wherein the user interface includes a blank menu option, the blank menu option permitting the input of a new information choice not offered by the set of valid user entries.

17. The KRS of claim 16 wherein all competed templates in which a blank menu option was selected are reviewed prior to transferring the fact and the ontology is modified to accept the new information choice if a reviewer of the fact approves the new information choice.

18. The KRS of claim 17 wherein the templates are completed by experts knowledgeable in the field represented by the ontology and the review of new information choices is conducted by the experts and knowledge base engineers.

19. Fact templates for use in a Knowledge Representation System (KRS) comprising:

at least one fact-type specific fact template, each fact-type specific fact template accepting one predefined fact type found in predetermined information sources wherein the template translates the fact type from one of a text-based, graphical and tabular language into a computational information language of the KRS; and

the at least one fact-type specific fact template having at least one user interface that constrains user data entry to one of a predetermined set of valid user entries that may be inserted into the fact template at the user interface.

20. The fact templates of claim 19 wherein scientists expert in the field defined by the predetermined information sources extract information from the predetermined information sources and complete the templates with the extracted information.

21. The fact templates of claim 20 wherein the user interface includes a user entry field that permits entry of new information not included in the predetermined set of valid user entries.

22. The fact templates of claim 21 wherein information entered into the user entry field automatically flags the completed fact template for review.

23. The fact templates of claim 22 wherein the ontology is modified to accept the new information if the fact template passes review.

24. A method of populating a Knowledge Representation System (KRS), comprising the steps of:

creating at least one template for the entry of a fact, wherein the at least one template is comprised of at least one user interface object that constrains user data entry based upon a set of valid entries, wherein the valid entries are determined by the structure of an ontology;

presenting the at least one template to at least one reviewer of information;

accepting a fact from the at least one reviewer, wherein the fact is entered into the at least one template as selections from the at least one user interface object; and

inserting the fact into the KRS.

25. The method of claim 24, wherein the at least one user interface object is a pull down menu.

26. The method of claim 24, wherein the valid entries for the at least one template conform to the organization and permitted content of the ontology.

27. The method of claim 24, wherein at least one template is modified if facts are found that cannot be accommodated by at least one existing template.

28. The method of claim 27, wherein the modification is to add at least one valid user entry to at least one menu in at least one template.

29. The method of claim 27, wherein the ontology is modified to accept new facts.

30. The method of claim 24, wherein content scientists knowledgeable in the field of knowledge represented by the ontology enter facts into the at least one template.

31. The method of claim 24, wherein the ontology is organized to store genomics information.

32. A computer-readable medium containing instructions that cause a processor to populate a knowledge representation system, by:

creating at least one template for the entry of a fact, wherein the at least one template is comprised of at least one user interface object that constrains user data entry based upon a set of valid entries, wherein the valid entries are determined by the structure of an ontology;

presenting the at least one template to at least one reviewer of information;

accepting a fact from the at least one reviewer, wherein the fact is entered into the at least one template as selections from the at least one user interface object; and

inserting the fact into a knowledge base according to the structure of the ontology.

33. The computer-readable medium of claim 32, wherein the options available on at least one template conform to the organization and permitted content of the ontology.

34. The computer-readable medium of claim 32, wherein at least one template is modified if facts are found that cannot be accommodated by at least one existing template.

35. The computer-readable medium of claim 34, wherein the modification is to add at least one selection to the at least user interface object in at least one template.

36. The method of claim 32, wherein the ontology is modified to accept new facts.

37. The method of claim 32, wherein content scientists knowledgeable in the field of knowledge represented by the ontology enter facts into the at least one template.

38. The method of claim 32, wherein the ontology is genomics information.