Systems and methods to index and search voice sites

• US 8,046,220 B2
• Filed: 11/28/2007
• Issued: 10/25/2011
• Est. Priority Date: 11/28/2007
• Status: Expired due to Fees

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Agarwal, Sheetal K., Kumar, Arun, Mukherjea, Sougata, Nanavati, Amit Anil, Rajput, Nitendra, Chakraborty, Dipanjan
• Primary Examiner(s): MCFADDEN, SUSAN IRIS
• Application Number:
  US11/946,122
• Publication Number:
  US 20090138262A1
• Time in Patent Office: 1,427 Days
• Field of Search: 704/235, 379/88.17
• US Class Current: 704/235
• CPC Class Codes: G06F 16/68 Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
• CPC Class Codes: G06F 16/685 using automatically derived transcript of audio data, e.g. lyrics speech recognition G10L15/00
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: H04M 2201/40 using speech recognition speech recognition per se G10L15/00
• CPC Class Codes: H04M 2201/60 Medium conversion
• CPC Class Codes: H04M 3/4938 comprising a voice browser which renders and interprets, e.g. VoiceXML

Abstract

A method includes crawling and indexing voice sites and storing results in an index; receiving a search request in voice from a user via a telephone; performing speech recognition on the voice search request and converting the request from voice to text; parsing the query; and performing a search on the index and ranking the search results. Search results may be filtered based on attributes such as location and context. Filtered search results may be presented to the user in categories to enable easy voice browsing of the search results by the user. Computer program code and systems are also provided.

Claims

1. A system comprising:

circuitry configured to crawl voice sites;

circuitry configured to index voice sites and store results in a voice site index, wherein each voice site comprises a voice driven application that includes one or more voice pages hosted on servers or computers in a telecom infrastructure;

circuitry configured to receive a search request in voice from a user via a telephone;

a speech to text generator configured to perform speech recognition on the voice search request and convert the request from voice to text;

a parser configured to parse the text search request;

circuitry configured to perform, in response to the parsed search request, a search on the voice site index and rank the search results;

circuitry configured to filter the search results based on at least one of location, emotion and context; and

a voice interface configured to present the filtered search results to the user.

2. A system in accordance with claim 1, wherein the voice interface provides the filtered results to the user in categories, for efficient voice browsing of the filtered search results.

3. A system in accordance with claim 1, wherein the circuitry configured to crawl is defined by a server.

4. A system in accordance with claim 3, wherein the circuitry configured to receive a search request in voice is defined by a server.

5. A system in accordance with claim 4, wherein the first mentioned server is different from the second mentioned server.

6. A system in accordance with claim 1, wherein the circuitry configured to index is defined by a server.

7. A system in accordance with claim 1, wherein the circuitry configured to index is configured to index voice sites using information contained in at least one of text prompts, audio prompts, meta information, and grammars from voice sites.

8. A system in accordance with claim 1, wherein the parser is defined by a server.

9. A system in accordance with claim 1, wherein the parser is configured to parse a search query to identify different types of terms in the query, the types including at least one of location, emotion and context, in addition to content terms.

10. A system in accordance with claim 1, wherein the circuitry configured to crawl is defined by a computer.

11. A method comprising:

crawling and indexing voice sites and storing results in a voice site index, wherein each voice site comprises a voice driven application that includes one or more voice pages hosted on servers or computers in a telecom infrastructure;

receiving a search request in voice from a user via a telephone;

performing speech recognition on the voice search request and converting the request from voice to text;

parsing the text search request;

performing, in response to the parsed search request, a search on the voice site index and ranking the search results;

filtering search results based on at least one of location, emotion and context; and

presenting the filtered search results to the user in categories to enable easy voice browsing of the search results by the user.

12. A method in accordance with claim 11, further configured to receive a search request from a user via the World Wide Web.

13. A method in accordance with claim 11, wherein the crawling is performed by extracting source code through an API and comprises extracting grammar items from voice sites, wherein grammar items are possible responses to voice prompts that the voice site expects.

14. A method in accordance with claim 11, wherein the crawling comprises extracting text prompts from voice sites.

15. A method in accordance with claim 11, wherein the crawling comprises extracting audio prompts from voice sites and converting the voice to text.

16. A method in accordance with claim 11, wherein the crawling is performed by extracting source code through an API and comprises extracting meta information from voice sites.

17. A method in accordance with claim 11, wherein the parsing comprises identifying at least one of location, language, and emotion.

18. A non-transitory memory encoded with instructions that, when executed by a computer, causes the computer to:

receive a voice search request from a user via a telephone;

perform speech recognition on the voice search request and convert the request from voice to text;

parse the text search request;

perform, in response to the parsed search request, a search on an index of voice sites and rank the search results, wherein each voice site comprises a voice driven application that includes one or more voice pages hosted on servers or computers in a telecom infrastructure; and

present at least one relevant search result to the user such that the user can audibly hear the results over the telephone.

19. A memory in accordance with claim 18, wherein the parsing comprises looking for terms indicating at least one of location, language, and emotion, in the search request.

20. A memory in accordance with claim 18, wherein presenting at least one relevant search result comprises providing search results in categories.