User Query History Expansion for Improving Language Model Adaptation

• US 20130080162A1
• Filed: 09/23/2011
• Published: 03/28/2013
• Est. Priority Date: 09/23/2011
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Chang, Shuangyu, Levit, Michael, Buntschuh, Bruce Melvin
• Granted Patent: US 9,129,606 B2
• Time in Patent Office: Days
• US Class Current: 704/235
• CPC Class Codes: G06F 16/24575 using context
• CPC Class Codes: G06F 16/9024 Graphs; Linked lists G06F16/9027 takes precedence
• CPC Class Codes: G06F 16/90324 using system suggestions
• CPC Class Codes: G06F 16/90332 Natural language query formulation or dialogue systems
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G10L 15/18 using natural language modelling
• CPC Class Codes: G10L 15/34 Adaptation of a single recogniser for parallel processing, e.g. by use of multiple processors or cloud computing

Abstract

Query history expansion may be provided. Upon receiving a spoken query from a user, an adapted language model may be applied to convert the spoken query to text. The adapted language model may comprise a plurality of queries interpolated from the user'"'"'s previous queries and queries associated with other users. The spoken query may be executed and the results of the spoken query may be provided to the user.

Claims

1. A method for providing a query history expansion, the method comprising:

receiving a spoken query from a user;

applying an adapted language model to convert the spoken query to text;

executing the spoken query; and

providing at least one result of the spoken query to the user.

2. The method of claim 1, wherein the adapted language model is specific to the user.

3. The method of claim 1, wherein the adapted language model is shared among a plurality of users.

4. The method of claim 3, further comprising grouping the plurality of users according to a similarity in at least one previous query.

5. The method of claim 3, further comprising grouping the plurality of users according to at least one common characteristic shared among the plurality of users.

6. The method of claim 5, wherein the at least one common characteristic comprises at least one of the following:

a common organization association, a common demographic characteristic, a common spoken language, and a common geographic home area.

7. The method of claim 1, wherein the adapted language model comprises at least one previous query by the user and at least one expanded query associated with at least one second user.

8. The method of claim 7, wherein the at least one previous query of the user further comprises a same previous query of the at least one second user.

9. The method of claim 1, wherein the adapted language model comprises at least one expanded query associated with a clicked web link, wherein the clicked web link is also associated with at least one previous query of the user.

10. The method of claim 1, wherein the adapted language model comprises a plurality of expanded queries, wherein each of the plurality of expanded queries comprise a lexical similarity to at least one previous query of the user.

11. The method of claim 1, wherein the adapted language model comprises a plurality of expanded queries, wherein each of the plurality of expanded queries are associated with a search domain associated with at least one previous query of the user.

12. A system for providing a query history expansion, the system comprising:

a memory storage; and

a processing unit coupled to the memory storage, wherein the processing unit is operable to;

receive a plurality of previous queries associated with a user,

retrieve a second plurality of previous queries associated with a plurality of other users,

interpolate at least a subset of the second plurality of previous queries with the plurality of previous queries associated with the user,

create an adapted language model according to the interpolated queries;

receive a new query from the user, and

convert the new query into a text string according to the adapted language model.

13. The system of claim 12, wherein the adapted language model is shared among a second plurality of other users.

14. The system of claim 13, wherein the user and the second plurality of other users are each associated with at least one common characteristic.

15. The system of claim 14, wherein the at least one common characteristic comprises at least one of the following:

a common organization association, a common demographic characteristic, a common spoken language, and a common geographic home area.

16. The system of claim 12, wherein the at least a subset of the second plurality of previous queries are selected according to a lexical similarity to at least one of the plurality of previous queries associated with the user.

17. The system of claim 12, wherein the processing unit is further operative to:

plot a query-click graph of the plurality of previous queries associated with the user and the second plurality of previous queries associated with the plurality of other users; and

select the at least a subset of the second plurality of previous queries for interpolation according to a threshold distance between each of the second plurality of previous queries and at least one query of the plurality of previous queries associated with the user.

18. The system of claim 12, wherein the at least a subset of the second plurality of previous queries are selected according to a threshold value of overlap of query results associated with each of the second plurality of previous queries and at least one of the plurality of previous queries associated with the user.

19. The system of claim 12, wherein the processing unit is further operative to:

annotate each of the second plurality of queries and each of the plurality of previous queries associated with the user; and

select the at least a subset of the second plurality of previous queries for interpolation according to a common annotation between at least one of the second plurality of previous queries and at least one of the plurality of queries associated with the user.

20. A computer-readable medium which stores a set of instructions which when executed performs a method for providing a query history expansion, the method executed by the set of instructions comprising:

collecting a plurality of previous search queries associated with a user, wherein the plurality of previous search queries comprises at least one of the following;

a previously spoken query, a previously typed query, and a previously selected query;

creating an adaptation data set associated with the user according to the plurality of previous queries;

expanding the adaptation data set by adding a plurality of historical queries from a plurality of users, wherein the plurality of users share at least one common characteristic with the user;

expanding the adaptation data set by adding a second plurality of queries, wherein each of the second the plurality of queries comprises a lexical similarity to at least one of the previous search queries associated with the user;

expanding the adaptation data set by adding a third plurality of queries, wherein each of the third plurality of historical queries is associated with a web link result, wherein each web link result is further associated with at least one of the plurality of previous search queries;

interpolating each of the plurality of previous search queries, the plurality of historical queries, the second plurality of queries, and the third plurality of queries into an adaptation data set for a personalized language model;

receiving a new search query spoken by the user;

converting the search query to text according to the personalized language model;

executing the converted query; and

providing at least one result of the executed query to the user.