Universally tagged frequent call-routing user queries as a knowledge base for reuse across applications

• US 8,406,384 B1
• Filed: 01/18/2012
• Issued: 03/26/2013
• Est. Priority Date: 01/18/2012
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Andreevskaia, Alina, Tremblay, Ral
• Primary Examiner(s): Nguyen, Joseph J
• Application Number:
  US13/352,640
• Time in Patent Office: 433 Days
• Field of Search: 379/67.1, 379/88.05, 379/88.06, 704/9, 704/257, 704/270.1, 704/275
• US Class Current: 379/88.05
• 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: G10L 15/1822 Parsing for meaning understanding
• CPC Class Codes: H04M 2201/40 using speech recognition speech recognition per se G10L15/00
• CPC Class Codes: H04M 2203/2061 Language aspects
• CPC Class Codes: H04M 2207/20 hybrid systems
• CPC Class Codes: H04M 2242/12 Language recognition, selection or translation arrangements
• CPC Class Codes: H04M 3/4936 Speech interaction details speech recognition per se G10L15/00
• CPC Class Codes: H04M 3/51 Centralised call answering arrangements requiring operator intervention , e.g. call or contact centers for telemarketing

Abstract

A computer-implemented method is described for developing query tags for classification of user queries to a call routing application. Multiple user query corpuses are accessed that contain user queries from multiple call routing applications in multiple different vertical domains. A set of frequent user queries are selected which appear in multiple different query corpuses in multiple different vertical domains. From these are developed frequent query tags for semantic classification of the frequent user queries. The frequent query tags are then stored in a query tag database.

Claims

1. A computer-implemented method employing at least one hardware based computer processor to develop query tags for classification of user queries to a call routing application, the method comprising:

accessing a plurality of user query corpuses containing user queries from a plurality of call routing applications in a plurality of different vertical domains;

selecting a set of frequent user queries that appear in a plurality of different query corpuses in a plurality of different vertical domains;

developing frequent query tags for semantic classification of the frequent user queries; and

storing the frequent query tags in a query tag database;

wherein the user queries and the query tags are in a first language, and the method further comprises;

automatically translating the user queries into a second language; and

storing the translated user queries with the frequent query tags in a call routing database for a call routing application in the second language.

2. A method according to claim 1, further comprising:

accessing the plurality of user query corpuses;

selecting a set of common user queries that appear in a plurality of different query corpuses in a single common vertical domain;

developing common query tags for semantic classification of the common user queries; and

storing the common query tags in the query tag database.

3. A method according to claim 2, further comprising:

accessing the plurality of user query corpuses;

selecting in an computer process a set of widespread user queries not previously selected;

developing widespread query tags for semantic classification of the widespread user queries; and

storing the widespread query tags in the query tag database.

4. A method according to claim 1, further comprising:

automatically assigning query tags from the query tag database to user queries from a new call routing application.

5. A method according to claim 4, further comprising:

establish a call routing ontology for the new application organizing the query tags for the new application in a class hierarchy reflecting associated semantic concepts.

6. A method according to claim 4, wherein the new call routing application is in a vertical domain represented in the user query corpuses, and wherein the assigned query tags are obtained by intersecting the query tags in the query tag database with the user queries in the user query corpuses in the vertical domain of the new call routing application.

7. A method according to claim 1, further comprising:

using the stored query tags in a clustering algorithm for a call steering application.

8. A method according to claim 7, further comprising:

using the clustering algorithm in a user tagging interface for a call steering application.

9. A method according to claim 1, further comprising:

using the stored query tags to build a statistical language model for audio clustering in an automatic speech recognition application.

10. A method according to claim 1, further comprising:

using the stored query tags as an input for injection of soft knowledge in a call steering application.

11. A system for developing query tags for classification of user queries to a call routing application, the system comprising:

a plurality of user query corpuses containing user queries from a plurality of call routing applications in a plurality of different vertical domains;

a query tag selection module for accessing the user query corpuses to select a set of frequent user queries that appear in a plurality of different query corpuses in a plurality of different vertical domains to develop frequent query tags for semantic classification of the frequent user queries; and

a query tag database for storing the frequent query tags;

wherein the user queries and the query tags are in a first language, and the system further automatically translates the user queries into a second language and stores the translated user queries with the frequent query tags in a call routing database for a call routing application in the second language.

12. A system according to claim 11, wherein the query tag selection module further accesses the user query corpuses to select a set of common user queries that appear in a plurality of different query corpuses in a single common vertical domain to develop common query tags for semantic classification of the common user queries and store the common query tags in the query tag database.

13. A system according to claim 12, wherein the query tag selection module further accesses the user query corpuses to select a set of widespread user queries that appear in a plurality of different query corpuses in a single widespread vertical domain to develop widespread query tags for semantic classification of the widespread user queries and store the widespread query tags in the query tag database.

14. A system according to claim 11, wherein the query tags from the query tag database are automatically assigned to user queries from a new call routing application.

15. A system according to claim 14, further comprising:

a call routing ontology for the new application that organizes the query tags for the new application in a class hierarchy reflecting associated semantic concepts.

16. A system according to claim 14, wherein the new call routing application is in a vertical domain represented in the user query corpuses, and wherein the assigned query tags are obtained by intersecting the query tags in the query tag database with the user queries in the user query corpuses in the vertical domain of the new call routing application.

17. A system according to claim 11, further comprising:

a clustering algorithm module of a call steering application that uses the stored query tags.

18. A system according to claim 17, further comprising:

a user tagging interface of a call steering application that uses the clustering algorithm.

19. A system according to claim 11, further comprising:

a statistical language model for audio clustering in an automatic speech recognition application derived from the stored query tags.

20. A system according to claim 11, further comprising:

a soft knowledge input of a call steering application that receives the stored query tags.