METHOD AND SYSTEM FOR TRANSLATION OF CROSS-LANGUAGE QUERY REQUEST AND CROSS-LANGUAGE INFORMATION RETRIEVAL

• US 20080235202A1
• Filed: 02/25/2008
• Published: 09/25/2008
• Est. Priority Date: 03/19/2007
• Status: Abandoned Application

• Current Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Original Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Inventors: Zhu, Jiang, WANG, Haifeng
• Application Number:
  US12/036,584
• Publication Number:
  US 20080235202A1
• Time in Patent Office: Days
• US Class Current: 707/4
• CPC Class Codes: G06F 16/3337 Translation of the query language, e.g. Chinese to English

Abstract

The present invention provides a method and apparatus for translation of a cross-language query request as well as a cross-language information retrieval method and system. The method for translation of a cross-language query request comprises: translating the cross-language query request from source language into a target language respectively with a plurality of different machine translation systems to obtain a plurality of translations in said target language of the cross-language query request; and constructing a target language query request corresponding to the cross-language query request based on said plurality of translations in said target language of the cross-language query request. The present invention constructs a target language query request by merging translations of cross-language query request generated by a plurality of different machine translation systems and hence improves the retrieval performance of cross-language information retrieval system.

Claims

1. A method for translation of a cross-language query request, comprising:

translating the cross-language query request from source language into a target language respectively with a plurality of different machine translation systems to obtain a plurality of translations in said target language of the cross-language query request; and

constructing a target language query request corresponding to the cross-language query request based on said plurality of translations in said target language of the cross-language query request.

2. The method for translation of a cross-language query request according to claim 1, wherein said step of constructing a target language query request further comprises:

merging said plurality of translations in said target language of the cross-language query request to form a query word list;

computing a weight for each query word in the query word list; and

constructing a target language query request corresponding to the cross-language query request based on the query word list and the weight of each query word in the query word list.

3. The method for translation of a cross-language query request according to claim 2, wherein said step of computing a weight for each query word in the query word list further comprises:

calculating a Translation Confidence for each of said plurality of translations in said target language of the cross-language query request; and

using the Translation Confidence of each of said plurality of translations in said target language of the cross-language query request in the computing of the weight for each query word in the query word list.

4. The method for translation of a cross-language query request according to claim 3, wherein said step of calculating a Translation Confidence further comprises:

acquiring a Translation Quality Score of each of the plurality of different machine translation systems;

calculating a LM Confidence for each of said plurality of translations in said target language of the cross-language query request with a language model; and

for each of said plurality of translations in said target language of the cross-language query request, combining the Translation Quality Score of the machine translation system generating the translation in said target language and the LM Confidence of the translation in said target language to obtain the Translation Confidence thereof.

5. The method for translation of a cross-language query request according to claim 4, wherein said step of combining the Translation Quality Score of the machine translation system generating the translation in said target language and the LM Confidence of the translation in said target language further comprises:

multiplying the Translation Quality Score of the machine translation system generating the translation in said target language by the LM Confidence of the translation in said target language.

6. The method for translation of a cross-language query request according to claim 4, wherein the Translation Quality Score of each of the plurality of different machine translation systems is previously generated by evaluating translation quality with respect to the machine translation system.

7. The method for translation of a cross-language query request according to any one of claims 3˜

6, wherein said step of using the Translation Confidence of each of said plurality of translations in said target language of the cross-language query request in the computing of the weight for each query word in the query word list further comprises;

using the Translation Confidence of each of said plurality of translations in said target language of the cross-language query request in the computing of the weighted term frequency for each query word in the query word list.

8. The method for translation of a cross-language query request according to any one of claims 3˜

6, wherein said step of using the Translation Confidence of each of said plurality of translations in said target language of the cross-language query request in the computing of the weight for each query word in the query word list further comprises;

computing the weight for each query word in the query word list using the Translation Confidence of each of said plurality of translations in said target language of the cross-language query request according to the following algorithm;

W_q,i=TF_q,i*IDF_i

where

$I<br><br><br><br>D<br><br><br><br>F_i=\log <br><br><br><br>\frac{D}{d_i},<br><br>{\mathrm{TF}}_{q,i}=\underset{i=1}{\overset{N}{\sum <br><br>}}<br><br>{\mathrm{TC}}_t*{\mathrm{freq}}_{t,i}$

wherein, W_q,i is the weight of query word i in the cross-language query request q;

TF_q,i is the weighted term frequency of query word i in the cross-language query request q;

IDF_i is the inverse document frequency of query word i;

D is the total number of documents;

d_i is the number of documents containing query word i;

freq_t,i is the occurrence times of query word i in the translation t in said target language of the cross-language query request q;

TC_t is the Translation Confidence of the translation t in said target language of the cross-language query request q.

9. The method for translation of a cross-language query request according to claim 1, wherein the target language query request is the set of query word-weight pairs respectively corresponding to a query word in the cross-language query request.

10. The method for translation of a cross-language query request according to claim 9, wherein the query word-weight pairs are in the form of <

query word;

weight>

.

11. A cross-language information retrieval method, comprising:

accepting a cross-language query request from a query user;

translating the cross-language query request from source language into a target language using the method for translation of a cross-language query request according to any one of the preceding claims 1˜

10 to generate a target language query request corresponding to the cross-language query request; and

retrieving documents in said target language meeting the target language query request from an information source.

12. The cross-language information retrieval method according to claim 11, further comprising:

presenting the documents in said target language meeting the target language query request to the query user.

13. An apparatus for translation of a cross-language query request, comprising:

a plurality of machine translation modules each configured to translate the cross-language query request from source language into a target language, thereby a plurality of translations in said target language of the cross-language query request are obtained; and

a target language query request construction module configured to construct a target language query request corresponding to the cross-language query request based on said plurality of translations in said target language of the cross-language query request.

14. The apparatus for translation of a cross-language query request according to claim 13, wherein the target language query request construction module further comprises:

a query word list formation module configured to merge said plurality of translations in said target language of the cross-language query request to form a query word list;

a weight computation module configured to compute a weight for each query word in the query word list; and

a query formulation generation module configured to generate a target language query formulation corresponding to the cross-language query request based on the query word list formed by the query word list formation module and the weight of each query word in the query word list computed by the weight computation module.

15. The apparatus for translation of a cross-language query request according to claims 13 or 14, wherein the target language query request construction module further comprises:

a Translation Confidence calculation module configured to calculate a Translation Confidence for each of the translations in said target language of the cross-language query request generated by said plurality of machine translation modules;

wherein the weight computation module uses the Translation Confidence of each of said plurality of translations in said target language calculated by the Translation Confidence calculation module in the computing of the weight for each query word in the query word list.

16. The apparatus for translation of a cross-language query request according to claim 15, wherein the Translation Confidence calculation module further comprises:

a Translation Quality evaluation module configured to evaluate translation quality for each of said plurality of machine translation modules to acquire a Translation Quality Score of the machine translation module; and

a LM Confidence calculation module configured to calculate a LM Confidence for each of the translations in said target language of the cross-language query request generated by said plurality of machine translation modules with a language model;

wherein the Translation Confidence calculation module, for each of said plurality of translations in said target language of the cross-language query request, multiplies the Translation Quality Score of the machine translation module generating the translation, which is evaluated by the Translation Quality evaluation module, by the LM Confidence of the translation in said target language, which is calculated by the LM Confidence calculation module, to obtain the Translation Confidence of the translation in said target language.

17. The apparatus for translation of a cross-language query request according to claim 15, wherein the weight computation module compute the weight for each query word in the query word list according to the following algorithm:

W_q,i=TF_q,i*IDF_i

where

$I<br><br><br><br>D<br><br><br><br>F_i=\log <br><br><br><br>\frac{D}{d_i},<br><br>{\mathrm{TF}}_{q,i}=\underset{i=1}{\overset{N}{\sum <br><br>}}<br><br>{\mathrm{TC}}_t*{\mathrm{freq}}_{t,i}$

wherein, W_q,i is the weight of query word i in the cross-language query request q;

TF_q,i is the weighted term frequency of query word i in the cross-language query request q;

IDF_i is the inverse document frequency of query word i;

D is the total number of documents;

d_i is the number of documents containing query word i;

freq_t,i is the occurrence times of query word i in the translation t in said target language of the cross-language query request q;

TC_t is the Translation Confidence of the translation tin said target language of the cross-language query request q.

18. A cross-language information retrieval system, comprising:

an user module configured to accept a cross-language query request from a query user and present retrieval result by the cross-language information retrieval system to the query user;

the apparatus for translation of a cross-language query request according to any one of claims 13˜

17 for translating the cross-language query request from source language into a target language to generate a target language query request corresponding to the cross-language query request; and

a retrieval module configured to retrieve documents in said target language meeting the target language query request from an information source.