Question answering system, data search method, and computer program

• US 7,526,474 B2
• Filed: 09/22/2005
• Issued: 04/28/2009
• Est. Priority Date: 03/14/2005
• Status: Expired due to Fees

• Current Assignee: Fuji Xerox Company Limited (Fujifilm Holdings Corporation)
• Original Assignee: Fuji Xerox Company Limited (Fujifilm Holdings Corporation)
• Inventors: Masuichi, Hiroshi, Yoshimura, Hiroki, Sugihara, Daigo, Ohkuma, Tomoko
• Primary Examiner(s): Woo; Isaac M
• Application Number:
  US11/232,379
• Publication Number:
  US 20060206481A1
• Time in Patent Office: 1,314 Days
• Field of Search: 707 1- 10, 707100-1041, 707200-206
• US Class Current: 1/1
• CPC Class Codes: G06F 16/3329 Natural language query formulation or dialogue systems
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching
• CPC Class Codes: Y10S 707/99937 Sorting
• CPC Class Codes: Y10S 707/99938 Concurrency, e.g. lock management in shared database
• CPC Class Codes: Y10S 707/99939 Privileged access

Abstract

A question answering system includes a question-sentence analysis unit, a search unit, an answer-candidate extraction unit, a check-expression generation unit, a check-expression search unit and an answer-candidate correction unit. The question-sentence analysis unit extracts a keyword applied to a search process from an input question sentence. The search unit executes the search process with using the extracted keyword. The answer-candidate extraction unit extracts answer candidates for the input question sentence based on a search result. The check-expression generation unit analyzes the input question sentence to generate a check expression concatenating an analysis result and the extracted answer candidates. The check-expression search unit executes a search process based on the generated check expression. The answer-candidate correction unit selects an output answer to the input question from among answer the candidates based on a search result.

Claims

1. A question answering system comprising:

a processor;

an input device for receiving an input question sentence;

a question-sentence analysis unit controlled by the processor which is configured to extract a keyword to be applied to a search process from the input question sentence;

a search unit controlled by the processor which is configured to execute the search process using the extracted keyword;

an answer-candidate extraction unit controlled by the processor which is configured to extract answer candidates for the input question sentence based on a search result of the search unit;

a check-expression generation unit controlled by the processor which is configured to analyze a structure of the input question sentence, an analysis result of the check-expression generation unit including at least detection of a subject of the input question sentence, and to generate a check expression in which (i) at least the subject of the input question sentence of the analysis result of the check-expression generation unit and (ii) each of the answer candidates extracted by the answer-candidate extraction unit are concatenated;

a check-expression search unit controlled by the processor which is configured to execute a search process based on the check expression generated by the check-expression generation unit; and

an answer-candidate correction unit controlled by the processor which is configured to select an output answer to the input question from among answer the candidates based on a search result of the check-expression search unit; and

an output device for displaying or storing the output answer.

2. The question answering system according to claim 1, wherein the check-expression generation unit is configured to execute a syntactic and semantic analysis process to analyze the input question sentence.

4. The question answering system according to claim 1, wherein the answer-candidate correction unit is configured to

calculate a number of hits corresponding to each answer candidate from the search result of the check-expression search unit and

select an answer candidate having the number of hits greater than a predetermined threshold value as the output answer to the input question.

5. The question answering system according to claim 1, wherein the answer-candidate extraction unit is configured to generate an answer candidate list ranking the answer candidates for the input question sentence in accordance with a predetermined score calculation algorithm.

3. A question answering system comprising:

a processor;

an input device for receiving an input question sentence;

a question-sentence analysis unit controlled by the processor which is configured to extract a keyword to be applied to a search process from the input question sentence;

a search unit controlled by the processor which is configured to execute the search process with using the extracted keyword;

an answer-candidate extraction unit controlled by the processor which is configured to extract answer candidates for the input question sentence based on a search result of the search unit;

a check-expression generation unit controlled by the processor which is configured to analyze a structure of the input question sentence to generate a check expression concatenating an analysis result of the check-expression generation unit and the answer candidates extracted by the answer-candidate extraction unit;

a check-expression search unit which is configured to execute a search process based on the check expression generated by the check-expression generation unit;

an answer-candidate correction unit controlled by the processor which is configured to select an output answer to the input question from among answer the candidates based on a search result of the check-expression search unit; and

an output device for displaying or storing the output answer,

wherein the check-expression generation unit is configured to

execute a syntactic and semantic analysis process of the input question sentence to generate f-structure representing semantic information of the question sentence as a matrix structure of attribute-value,

generate f-structure, which corresponds to each answer candidate and concatenates each answer candidate extracted by the answer-candidate extraction unit with the generated f-structure or simplified f-structure resulting from simplifying the generated f-structure, and

generates the check expression based on the generated f-structure corresponding to each answer candidate.

6. A question answering system comprising:

a processor;

an input device for receiving an input question sentence;

a question-sentence analysis unit controlled by the processor which is configured to extract a keyword to be applied to a search process from the input question sentence;

a search unit controlled by the processor which is configured to execute the search process using the extracted keyword;

an answer-candidate extraction unit controlled by the processor which is configured to extract answer candidates for the input question sentence based on a search result of the search unit;

a check-expression generation unit controlled by the processor which is configured to analyze a structure of the input question sentence to generate a check expression concatenating an analysis result of the check-expression generation unit and the answer candidates extracted by the answer-candidate extraction unit;

a check-expression search unit controlled by the processor which is configured to execute a search process based on the check expression generated by the check-expression generation unit;

an answer-candidate correction unit controlled by the processor which is configured to select an output answer to the input question from among answer the candidates based on a search result of the check-expression search unit; and

an output device for displaying or storing the output answer,

wherein the check-expression generation unit is configured to

execute a syntactic and semantic analysis process of the input question sentence,

generate an f-structure representing semantic information of the sentence as a matrix structure of attribute-value as the analysis result, and

generate an f-structure, which corresponds to each answer candidate and concatenate each answer candidate extracted by the answer-candidate extraction unit with the generated f-structure or simplified f-structure resulting from simplifying the generated f-structure; and

the check-expression generation unit being configured to execute the search process based on the f-structure corresponding to each answer candidate.

7. A data search method comprising:

extracting a keyword to be applied to a first search process from an input question sentence;

executing the first search process with using the extracted keyword;

extracting answer candidates for the input question sentence based on a search result of the first search process;

executing an analysis of the input question sentence, an analysis result including at least detection of a subject of the input question sentence, to generate a check expression in which (i) at least the subject of the input question sentence of the analysis result and (ii) each of the answer candidates are concatenated;

executing a second search process based on the generated check expression; and

selecting an output answer to the input question from among the answer candidates based on a search result of the second search process.

8. The data search method according to claim 7, wherein the executing of the analysis of the input question sentence comprises:

executing a syntactic and semantic analysis process of the input question sentence.

9. The data search method according to claim 7, wherein the selecting of the output answer comprises:

calculating number of hits corresponding to each answer candidate from the search result of the second search process; and

selecting an answer candidate having the number of hits greater than a predetermined threshold value as the output answer to the input question.

10. The data search method according to claim 7, wherein the extracting of the answer candidates comprises generating an answer candidate list ranking the answer candidates for the input question sentence in accordance with a predetermined score calculation algorithm.

11. A data search method comprising:

extracting a keyword to be applied to a first search process from an input question sentence;

executing the first search process with using the extracted keyword;

extracting answer candidates for the input question sentence based on a search result of the first search process;

executing an analysis of the input question sentence to generate a check expression concatenating an analysis result and the extracted answer candidates;

executing a second search process based on the generated check expression; and

selecting an output answer to the input question from among the answer candidates based on a search result of the second search process;

wherein the executing of the analysis of the input question sentence comprises;

executing a syntactic and semantic analysis process of the input question sentence to generates f-structure representing semantic information of the question sentence as a matrix structure of attribute-value,

generating f-structure, which corresponds to each answer candidate and concatenates each answer candidate extracted by the answer-candidate extraction unit with the generated f-structure or simplified f-structure resulting from simplifying the generated f-structure, and

generating the check expression based on the generated f-structure corresponding to each answer candidate.

12. A data search method comprising:

extracting a keyword to be applied to a first search process from an input question sentence;

executing the first search process with using the extracted keyword;

extracting answer candidates for the input question sentence based on a search result of the first search process;

executing an analysis of the input question sentence to generate a check expression concatenating an analysis result and the extracted answer candidates;

executing a second search process based on the generated check expression; and

selecting an output answer to the input question from among the answer candidates based on a search result of the second search process;

wherein the executing of the analysis of the input question sentence comprises;

executing a syntactic and semantic analysis process of the input question sentence,

generating f-structure representing semantic information of the sentence as a matrix structure of attribute-value as the analysis result,

generating f-structure, which corresponds to each answer candidate and concatenates each extracted answer with the generated f-structure or simplified f-structure resulting from simplifying the generated f-structure, and

executing the second search process based on the f-structure corresponding to each answer candidate.

13. A computer program stored in a computer readable medium, the computer program causing a computer to execute a data search process comprising:

extracting a keyword to be applied to a first search process from an input question sentence;

executing the first search process with using the extracted keyword;

extracting answer candidates for the input question sentence based on a search result of the first search process;

executing an analysis of the input question sentence, an analysis result including at least detection of a subject of the input question sentence, to generate a check expression in which (i) at least the subject of the input question sentence of the analysis result and (ii) each of the answer candidates are concatenated;

executing a second search process based on the generated check expression; and

selecting an output answer to the input question from among the answer candidates based on a search result of the second search process.