Method and system for detecting user intentions in retrieval of hint sentences

• US 20040059730A1
• Filed: 09/19/2002
• Published: 03/25/2004
• Est. Priority Date: 09/19/2002
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Zhou, Ming
• Granted Patent: US 7,293,015 B2
• Time in Patent Office: Days
• US Class Current: 707/4
• CPC Class Codes: G06F 40/253 Grammatical analysis; Style critique
• CPC Class Codes: G06F 40/58 Use of machine translation, e.g. for multi-lingual retrieval, for server-side translation for client devices or for real-time translation
• CPC Class Codes: Y10S 707/99932 Access augmentation or optimizing
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access
• CPC Class Codes: Y10S 707/99936 Pattern matching access
• CPC Class Codes: Y10S 707/99943 Generating database or data structure, e.g. via user interface

Abstract

A method of constructing a confusion set database for use in detecting user query intentions includes obtaining a bilingual database having aligned word pairs in first and second languages. Second language word pairs in the bilingual database are aligned with corresponding correct translation first language word pairs. First language human translation word pairs corresponding to each of the second language word pairs in the bilingual database are obtained. Each first language human translation word pair for a particular second language word pair in the bilingual database is aligned with the correct translation first language word pair to define first language set pairs in the confusion set database. Methods, systems and computer readable medium for constructing the confusion set database and for retrieving sentences using the confusion set database are also disclosed.

Claims

1. A method of constructing a confusion set database for use in detecting user query intentions in the retrieval of sentences from a sentence database in response to the query, the method comprising:

obtaining a bilingual database having word pairs in first and second languages, wherein each of a plurality of second language word pairs in the bilingual database is aligned with a corresponding correct translation first language word pair;

obtaining first language human translation word pairs corresponding to each of the plurality of second language word pairs in the bilingual database;

aligning each first language human translation word pair for a particular second language word pair in the bilingual database with the correct translation first language word pair corresponding to the particular second language word pair to define first language set pairs in the confusion set database.

2. The method of claim 1, and further comprising:

removing from the confusion set database first language set pairs in which the first language human translation word pair and the correct translation first language word pair, defining the first language set pair, are the same.

3. The method of claim 1, wherein aligning each first language human translation word pair for a particular second language word pair with the correct translation first language word pair to define the first language set pairs in the confusion set database further comprises:

determining which first language set pairs have a first language human translation word pair which differs from its corresponding correct translation first language word pair; and

inserting into the confusion set database only those first language set pairs in which the first language human translation word pair differs from its corresponding correct translation first language word pair.

4. The method of claim 1, wherein obtaining the bilingual database having word pairs in the first and second languages includes:

obtaining a word and sentence aligned bilingual corpus.

5. The method of claim 1, wherein obtaining the bilingual database having word pairs in the first and second languages includes:

obtaining a word aligned bilingual dictionary.

6. The method of claim 1, wherein obtaining the first language human translation word pairs corresponding to each of the plurality of second language word pairs in the bilingual database further comprises:

performing human translation of each of the plurality of second language word pairs in the bilingual database into the first language human translation word pairs.

7. A computer-readable medium having computer-executable instructions for performing steps comprising:

obtaining a bilingual database having word pairs in first and second languages, wherein each of a plurality of second language word pair in the bilingual database is aligned with a corresponding correct translation first language word pair;

obtaining first language human translation word pairs corresponding to each of the plurality of second language word pairs in the bilingual database; and

aligning each first language human translation word pair for a particular second language word pair in the bilingual database with the correct translation first language word pair corresponding to the particular second language word pair to define first language set pairs in a confusion set database for use in detecting user query intentions in the retrieval of sentences from a sentence database in response to the query.

8. The computer readable medium of claim 7, and further having computer-executable instructions for performing further steps comprising:

removing from the confusion set database first language set pairs in which the first language human translation word pair and the correct translation first language word pair, defining the first language set pair, are the same.

9. The computer readable medium of claim 7, wherein aligning each first language human translation word pair for a particular second language word pair with the correct translation first language word pair to define the first language set pairs in the confusion set database further comprises:

determining which first language set pairs have a first language human translation word pair which differs from its corresponding correct translation first language word pair; and

inserting into the confusion set database only those first language set pairs in which the first language human translation word pair differs from its corresponding correct translation first language word pair.

10. A method of providing to a user sentences from a sentence database in response to a query, the method comprising:

receiving the query;

comparing word pairs in the query with human translation word pairs in a confusion set database to identify matches between word pairs in the query with human translation word pairs in the confusion set database, wherein the confusion set database includes a plurality of first language set pairs, and wherein each of the plurality of first language set pairs includes a correct translation first language word pair aligned with a first language human translation word pair;

for each word pair in the query which matched with a human translation word pair in the confusion set database, adding to the query the correct translation first language word pair corresponding to the matched human translation word pair to obtain an expanded query; and

retrieving at least one sentence from the sentence database in response to the expanded query.

11. The method of claim 10, wherein for each word pair in the query which matched with a human translation word pair in the confusion set database, adding to the query the correct translation first language word pair corresponding to the matched human translation word to obtain the expanded query further comprises:

replacing the word pair in the query with the correct translation first language word pair corresponding to the matched human translation word pair to obtain the expanded query.

12. The method of claim 10, wherein retrieving the at least one sentence from the sentence database in response to the expanded query further comprises:

retrieving at least one hint sentence from the sentence database in response to the expanded query.

13. A computer-readable medium having computer-executable instructions for performing steps comprising:

receiving the query;

comparing word pairs in the query with human translation word pairs in a confusion set database to identify matches between word pairs in the query with human translation word pairs in the confusion set database, wherein the confusion set database includes a plurality of first language set pairs, and wherein each of the plurality of first language set pairs includes a correct translation first language word pair aligned with a first language human translation word pair;

for each word pair in the query which matched with a human translation word pair in the confusion set database, adding to the query the correct translation first language word pair corresponding to the matched human translation word pair to obtain an expanded query; and

retrieving at least one sentence from a sentence database in response to the expanded query.

14. The computer-readable medium of claim 13, wherein for each word pair in the query which matched with a human translation word pair in the confusion set database, the step of adding to the query the correct translation first language word pair corresponding to the matched human translation word pair to obtain the expanded query further comprises:

replacing the word pair in the query with the correct translation first language word pair corresponding to the matched human translation word pair to obtain the expanded query.

15. The computer-readable medium of claim 13, wherein the step of retrieving the at least one sentence from the sentence database in response to the expanded query further comprises:

retrieving at least one hint sentence from the sentence database in response to the expanded query.

16. A system for retrieving sentences from a sentence database in response to a query, the system comprising:

an input component which receives the query as an input; and

a search engine coupled to the input component, the search engine comprising;

a comparison component which compares word pairs in the query to human translation word pairs in a confusion set database to identify matches between word pairs in the query with human translation word pairs in the confusion set database, wherein the confusion set database includes a plurality of first language set pairs, and wherein each of the plurality of first language set pairs includes a correct translation first language word pair aligned with a first language human translation word pair;

a query expansion component coupled to the comparison component, wherein for each word pair in the query which matched with a human translation word pair in the confusion set database, the query expansion component adds to the query the correct translation first language word pair corresponding to the matched human translation word pair to obtain an expanded query; and

a sentence retrieving component coupled to the query expansion component, the sentence retrieving component retrieving at least one sentence from a sentence database in response to the expanded query.

17. The system of claim 16, wherein the query expansion component is configured to, for each word pair in the query which matched with a human translation word pair in the confusion set database, replace the word pair in the query with the correct translation first language word pair corresponding to the matched human translation word pair to obtain the expanded query.

18. The system of claim 16, wherein the sentence retrieving component retrieves at least one hint sentence from the sentence database in response to the expanded query.