Methods and Systems for Query-Based Searching Using Spoken Input

Abstract

Systems and methods for query-based searching using spoken input are disclosed. In systems and methods according to embodiments of the invention, continuous speech natural language queries are accepted from a user using a client device. Speech processing tasks are divided between the client device and one or more server systems. Once user speech is recognized, the system searches one or more data repositories containing queries for at least one query that matches the recognized speech and returns information related to the query.

1 .- 28 . (canceled) 
   
   
       29 . A method for conducting a search using spoken input, comprising:
 accepting continuous user speech in the form of a natural language query using a client device;   processing the user speech on the client device using a first processing routine to form processed speech;   transmitting the processed speech to a server system using an application-level Internet protocol;   recognizing the processed speech on the server system using a second processing routine to form recognized speech;   searching one or more data repositories containing pluralities of queries to find at least one query that matches the recognized speech; and   returning information associated with the at least one query that matches the recognized speech to the client device.   
   
   
       30 . The method of  claim 29 , further comprising initiating the accepting after a button on the client device is depressed. 
   
   
       31 . The method of  claim 29 , wherein the processed speech comprises speech in compressed form. 
   
   
       32 . The method of  claim 31 , wherein the speech in compressed form comprises a plurality of extracted features characteristic of the user speech. 
   
   
       33 . The method of  claim 32 , wherein the extracted features comprise mel-frequency cepstral coefficients (MFCCs). 
   
   
       34 . The method of  claim 29 , wherein the Internet protocol comprises hypertext transfer protocol (HTTP). 
   
   
       35 . The method of  claim 29 , wherein transmitting the processed speech to the server system comprises transmitting the processed speech to the server system in streaming packets. 
   
   
       36 . A system for conducting a search using spoken input, comprising:
 a machine-readable medium encoded with an application program in the form of machine-readable instructions that, when executed, cause a client device to
 accept continuous user speech in the form of a natural language query using a client device, 
 process the user speech on the client device using a first processing routine to form processed speech, and 
 transmit the processed speech to a server system using an application-level Internet protocol; 
   a speech recognition engine that recognizes the processed speech to form recognized speech; and   a search engine that searches one or more data repositories containing pluralities of queries for at least one query corresponding to the recognized speech and returns information associated with the at least one query; and   a formatting and transmission routine that receives the processed speech and returns the information associated with the at least one query to the client device.   
   
   
       37 . The system of  claim 36 , wherein the speech recognition engine and the search engine reside in one or more server systems. 
   
   
       38 . The system of  claim 36 , wherein the information associated with the at least one query comprises an answer associated with the at least one query in the one or more data repositories. 
   
   
       39 . The system of  claim 36 , further comprising a natural language engine that analyzes a set of candidate responses to the at least one query that are returned by the search engine and determines the information associated with the at least one query that is returned. 
   
   
       40 . The system of  claim 39 , wherein the information associated with the at least one query comprises a single best response. 
   
   
       41 . The system of  claim 36 , wherein the processed speech comprises a spectral representation of the user speech. 
   
   
       42 . The system of  claim 36 , wherein the processed speech comprises speech in compressed form. 
   
   
       43 . The system of  claim 42 , wherein the speech in compressed form comprises a plurality of extracted features characteristic of the user speech. 
   
   
       44 . The system of  claim 43 , wherein the extracted features comprise mel-frequency cepstral coefficients (MFCCs). 
   
   
       45 . The system of  claim 36 , wherein the Internet protocol comprises hypertext transfer protocol (HTTP). 
   
   
       46 . A system for conducting a search using spoken input, comprising:
 a first speech processing circuit on a first device that
 accepts continuous user speech in the form of a natural language query using a client device, 
 processes the user speech on the client device using a first processing routine to form processed speech, and 
 transmit the processed speech to a second device using an application-level Internet protocol; 
   a second speech processing device on a second device that recognizes the processed speech to form recognized speech;   a search engine that searches one or more data repositories containing pluralities of queries for at least one query corresponding to the recognized speech and returns information associated with the at least one query; and   a formatting and transmission circuit that receives the processed speech and returns the information associated with the at least one query to the first device.   
   
   
       47 . The system of  claim 46 , further comprising a natural language engine that processes the recognized speech using both semantic decoding and statistical processing. 
   
   
       48 . The system of  claim 46 , wherein the information associated with the at least one query comprises at least one best answer to the at least one query. 
   
   
       49 . The system of  claim 46 , wherein the information associated with the at least one query comprises a set of candidate responses to the at least one query.