System and method for automated answering of natural language questions and queries

• US 7,209,876 B2
• Filed: 11/13/2002
• Issued: 04/24/2007
• Est. Priority Date: 11/13/2001
• Status: Active Grant

• Current Assignee: IoT Innovations LLC (Empire IP LLC)
• Original Assignee: Groove Unlimited, LLC
• Inventors: Miller, Michael J., Wantz, Leonard Jay
• Primary Examiner(s): ABEBE, DANIEL DEMELASH
• Application Number:
  US10/293,743
• Publication Number:
  US 20030093276A1
• Time in Patent Office: 1,623 Days
• Field of Search: 704/9, 707/4
• US Class Current: 704/9
• CPC Class Codes: G10L 15/18 using natural language modelling
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: Y10S 707/99934 Query formulation, input preparation, or translation

Abstract

A method for searching an information repository to find answers to natural language questions by generating an expected answer form based upon the form of the natural language question or query, searching for an answer meeting the specification of the expected answer form, and rendering an answer to a user.

Claims

1. A method for producing answers to a question or query issued to an information repository containing natural language data, executable in a computer system, comprising the steps of:

a. receiving a search question or query containing at least one variable expression, said at least one variable expression being representative of desired answer data that is to be searched for in said information repository;

b. initiating a search of said information repository for information containing terms present in said search question or query and receiving a first data set comprising information containing terms present in said search question or query;

c. searching said first data set for answer phrases present in syntactically similar form to said search question or query and containing answer data which may replace said variable expression to provide an answer to said search question or query, and retrieving answer phrases resulting from such search; and

d. modifying said answer phrases and providing output derived from said answer phrases in the form of one or more direct answers to said search question or query.

2. The method of claim 1, further comprising the step of:

(i) prior to step b, extracting key words and/or phrases from said search question or query, generating a key words and/or phrases search comprising said key words and/or phrases, and inputting said key words and/or phrases search to one or more search engines and/or search services.

3. The method of claim 1, further comprising the step of:

(i) prior to step a, receiving a natural language question, and converting said natural language question to said search question or query in the form of an expected answer form.

4. The method of claim 3, step (i) further comprising converting said natural language question to a generic question form, and converting said generic question form to said search question or query.

5. The method of claim 4, wherein said natural language question comprises generic question terms and substantive data terms, and said step of converting said natural language question to a generic question form further comprises replacing said substantive data terms with substantive data variable phrases.

6. The method of claim 5, wherein said step of converting said generic question form to said search question or query further comprises parsing said generic question form directly into an expected answer form having substantive data variable phrases, and substituting substantive data terms from said natural language question for substantive data variable phrases in said expected answer forms.

7. The method of claim 5, wherein said step of converting said generic question form to said search question or query further comprises comparing said generic question form to a listing of stored generic question forms having expected answer forms associated therewith, identifying a stored generic question form in said listing that matches said generic question form, retrieving expected answer forms associated with said identified generic question form and having substantive data variable phrases, and substituting substantive data terms from said natural language question for substantive data variable phrases in said expected answer forms.

8. The method of claim 7, said expected answer forms further comprising answer variable phrases representative of answer data to be retrieved, step c further comprising locating natural language phrases in said first data set containing said substantive data terms, and containing answer data arranged in said natural language phrases with respect to said substantive data terms so as to syntactically match said expected answer forms.

9. The method of claim 5, further comprising the step of:

(ii) prior to step b, extracting key words and/or phrases from said search question or query, generating a key words and/or phrases search comprising said key words and/or phrases, and inputting said key words and/or phrases search to one or more search engines and/or search services.

10. The method of claim 9, wherein said key words and/or phrases further comprise said substantive data terms in said natural language question.

11. The method of claim 1, step d further comprising highlighting a portion of said direct answers containing said data replacing said variable expression.

12. The method of claim 1, step d further comprising ranking said output based upon at least one of (i) redundant occurrence of an answer phrase, and (ii) length of said data replacing said variable expression.

13. A method for producing answers to a question or query, executable in a computer system, comprising the steps of:

a. receiving a search question or query comprising data indicative of a subject to be searched;

b. initiating a key words and/or phrases search of an information repository for data containing terms present in said search question or query, and receiving a first data set comprising data from said information repository containing terms present in said search question or query;

c. retrieving answer phrases from said first data set having a syntactic form matching an expected answer to said search question or query; and

d. providing output in the form of one or more of said answer phrases.

14. The method of claim 13, further comprising the step of:

(i) prior to step b, constructing a new search query comprising key words and/or phrases including said data indicative of a subject to be searched, said new search query configured as input to an information retrieval engine executable in a computer system.

15. The method of claim 13, further comprising the step of:

(i) prior to step b, extracting key words and/or phrases from said search question or query, generating a key words and/or phrases search comprising said key words and/or phrases, and inputting said key words and/or phrases search to a search engine.

16. The method of claim 13, further comprising the step of:

(i) prior to step a, receiving a natural language question, and converting said natural language question to said search question or query in the form of an expected answer form.

17. The method of claim 16, further comprising converting said natural language question to a generic question form, and converting said generic question form to said search question or query.

18. The method of claim 17, wherein said natural language question comprises generic question terms and substantive data terms, and said step of converting said natural language question to a generic question form further comprises replacing said substantive data terms with substantive data variable phrases.

19. The method of claim 18, wherein said step of converting said generic question form to said search question or query further comprises parsing said generic question form directly into an expected answer form having substantive data variable phrases, and substituting substantive data terms from said natural language question for substantive data variable phrases in said expected answer forms.

20. The method of claim 18, wherein said step of converting said generic question form to said search question or query further comprises comparing said generic question form to a listing of stored generic question forms having expected answer forms associated therewith, identifying a stored generic question form in said listing that matches said generic question form, retrieving expected answer forms associated with said identified generic question form and having substantive data variable phrases, and substituting substantive data terms from said natural language question for substantive data variable phrases in said expected answer forms.

21. The method of claim 20, said expected answer forms further comprising answer variable phrases representative of answer data to be retrieved, step c further comprising locating natural language phrases in said first data set containing said substantive data terms, and containing answer data arranged in said natural language phrases with respect to said substantive data terms so as to syntactically match said expected answer forms.

22. The method of claim 18, further comprising the step of:

(ii) prior to step b, extracting key words and/or phrases from said search question or query, generating a key words and/or phrases search comprising said key words and/or phrases, and inputting said key words and/or phrases search to one or more search engines and/or search services.

23. The method of claim 22, wherein said key words and/or phrases further comprise said substantive data terms in said natural language question.

24. The method of claim 13, step d further comprising modifying said answer phrases and providing output derived from and including said answer phrases, said output being in the form of one or more direct answers to said search question or query.

25. A method for processing a natural language question or query, executable in a computer system, comprising the steps of:

a. receiving a natural language question or query comprising generic question terms and substantive data terms;

b. converting said natural language question or query to a generic question form;

c. converting said generic question form to an expected answer form;

d. initiating a key words and/or phrases search of an information repository containing natural language data for information containing key words and/or phrases present in said key words and/or phrases search, and receiving a results set comprising information containing key words and/or phrases present in said key words and/or phrases search;

e. filtering said results set to retrieve answer phrases syntactically matching said expected answer form; and

f. formatting said answer phrases as direct answers to said natural language question or query, and producing output in the form of formatted answer phrases.

26. The method of claim 25, step b further comprising replacing said substantive data terms in said natural language question or query with substantive data variable phrases.

27. The method of claim 26, step c further comprising parsing said generic question form directly into an expected answer form having substantive data variable phrases, and substituting substantive data terms from said natural language question for substantive data variable phrases in said expected answer forms.

28. The method of claim 26, step c further comprising comparing said generic question form to a listing of stored generic question forms having expected answer forms associated therewith, identifying a stored generic question form in said listing that matches said generic question form, retrieving expected answer forms associated with said identified generic question form and having said substantive data variable phrases, and substituting substantive data terms from said natural language question or query for substantive data variable phrases in said expected answer forms.

29. The method of claim 28, step d further comprising, prior to initiating said key words and/or phrases search, extracting said substantive data terms from said expected answer forms and generating a key words and/or phrases search query comprising the extracted substantive data terms.

30. The method of claim 28, step e further comprising locating natural language phrases in said results set containing said substantive data terms, and containing answer data arranged in said natural language phrases with respect to said substantive data terms so as to syntactically match said expected answer form.

31. The method of claim 25, step f further comprising modifying said answer phrases to provide output derived from and including said answer phrases, said output being in the form of one or more direct answers to said natural language question or query.

32. A system for producing answers to a question or query comprising:

a server computer hosting a data search service accessible via client computers to a plurality of potential users, said server computer having computer readable program code means embodied therein and means for communicating with an information repository containing natural language data, said computer readable program code means further comprising;

a. program code means for converting a natural language question or query comprising generic question terms and substantive data terms into an expected answer form comprising said substantive data terms and answer variable phrases;

b. program code means for extracting said substantive data terms from said expected answer form, generating a key words and/or phrases search query comprising the extracted substantive data terms, and initiating a key words and/or phrases search of said information repository using said key words and/or phrases search query;

c. program code means for filtering results of said key words and/or phrases search to retrieve answer phrases syntactically matching said expected answer form; and

d. program code means for formatting said answer phrases as direct answers to said natural language question or query, and for producing output in the form of formatted answer phrases.