System and Method of Lattice-Based Search for Spoken Utterance Retrieval

Abstract

A system and method are disclosed for retrieving audio segments from a spoken document. The spoken document preferably is one having moderate word error rates such as telephone calls or teleconferences. The method comprises converting speech associated with a spoken document into a lattice representation and indexing the lattice representation of speech. These steps are performed typically off-line. Upon receiving a query from a user, the method further comprises searching the indexed lattice representation of speech and returning retrieved audio segments from the spoken document that match the user query.

We claim: 
     
         1 . A method comprising:
 indexing a lattice representation of speech in a spoken document, to yield an indexed lattice representation;   searching, via a processor, the indexed lattice representation based on a query associated with the spoken document, to yield one or more search results; and   based on the one or more search results, returning one or more audio segments from the spoken document which correspond to the query.   
     
     
         2 . The method of  claim 1 , further comprising:
 receiving the query from a user via a device.   
     
     
         3 . The method of  claim 1 , further comprising:
 searching a word index of the spoken document using the query, to yield the one or more search results.   
     
     
         4 . The method of  claim 1 , further comprising:
 searching a sub-word index of the spoken document using the query, to yield the one or more search results.   
     
     
         5 . The method of  claim 3 , wherein searching the word index of the spoken document using the query further comprises searching the word index when the query is in-vocabulary. 
     
     
         6 . The method of  claim 4 , wherein searching the sub-word index of the spoken document using the query further comprises searching the sub-word index when the query is out-of-vocabulary. 
     
     
         7 . The method of  claim 1 , wherein the query comprises at least one query word. 
     
     
         8 . A system comprising:
 a processor; and   a computer-readable storage medium having instructions stored which, when executed by the processor, cause the processor to perform operations comprising:   indexing a lattice representation of speech in a spoken document, to yield an indexed lattice representation;   searching the indexed lattice representation based on a query associated with the spoken document, to yield one or more search results; and   based on the one or more search results, returning one or more audio segments from the spoken document which correspond to the query.   
     
     
         9 . The system of  claim 8 , wherein the computer-readable storage medium stores additional instructions stored which, when executed by the processor, cause the processor to perform operations further comprising:
 receiving the query from a user via a device.   
     
     
         10 . The system of  claim 8 , wherein the computer-readable storage medium stores additional instructions stored which, when executed by the processor, cause the processor to perform operations further comprising:
 searching a word index of the spoken document using the query, to yield the one or more search results.   
     
     
         11 . The system of  claim 8 , wherein the computer-readable storage medium stores additional instructions stored which, when executed by the processor, cause the processor to perform operations further comprising:
 searching a sub-word index of the spoken document using the query, to yield the one or more search results.   
     
     
         12 . The system of  claim 10 , wherein searching the word index of the spoken document using the query further comprises searching the word index when the query is in-vocabulary. 
     
     
         13 . The system of  claim 11 , wherein searching the sub-word index of the spoken document using the query further comprises searching the sub-word index when the query is out-of-vocabulary. 
     
     
         14 . The system of  claim 8 , wherein the query comprises at least one query word. 
     
     
         15 . A computer-readable storage device having instructions stored which, when executed by a processor, cause the processor to perform operations comprising:
 indexing a lattice representation of speech in a spoken document, to yield an indexed lattice representation;
 searching the indexed lattice representation based on a query associated with the spoken document, to yield one or more search results; and 
 based on the one or more search results, returning one or more audio segments from the spoken document which correspond to the query. 
   
     
     
         16 . The computer-readable storage device of  claim 15 , wherein the computer-readable storage device stores additional instructions stored which, when executed by the processor, cause the processor to perform operations further comprising:
 receiving the query from a user via a device.   
     
     
         17 . The computer-readable storage device of  claim 15 , wherein the computer-readable storage device stores additional instructions stored which, when executed by the processor, cause the processor to perform operations further comprising:
 searching a word index of the spoken document using the query, to yield the one or more search results.   
     
     
         18 . The computer-readable storage device of  claim 15 , wherein the computer-readable storage device stores additional instructions stored which, when executed by the processor, cause the processor to perform operations further comprising:
 searching a sub-word index of the spoken document using the query, to yield the one or more search results.   
     
     
         19 . The computer-readable storage device of  claim 17 , wherein searching the word index of the spoken document using the query further comprises searching the word index when the query is in-vocabulary. 
     
     
         20 . The computer-readable storage device of  claim 18 , wherein searching the sub-word index of the spoken document using the query further comprises searching the sub-word index when the query is out-of-vocabulary.