Automated Extraction of Semantic Content and Generation of a Structured Document from Speech

Abstract

Techniques are disclosed for automatically generating structured documents based on speech, including identification of relevant concepts and their interpretation. In one embodiment, a structured document generator uses an integrated process to generate a structured textual document (such as a structured textual medical report) based on a spoken audio stream. The spoken audio stream may be recognized using a language model which includes a plurality of sub-models arranged in a hierarchical structure. Each of the sub-models may correspond to a concept that is expected to appear in the spoken audio stream. Different portions of the spoken audio stream may be recognized using different sub-models. The resulting structured textual document may have a hierarchical structure that corresponds to the hierarchical structure of the language sub-models that were used to generate the structured textual document.

1 . A method comprising steps of:
 (A) identifying a probabilistic language model including a plurality of probabilistic language models associated with a plurality of sub-structures of a document; and   (B) using a speech recognition decoder to apply the probabilistic language model to a spoken audio stream to produce a document including content organized into the plurality of sub-structures, wherein the content in each of the plurality of sub-structures is produced by recognizing speech using the probabilistic language model associated with the sub-structure.   
     
     
         2 . The method of  claim 1 , wherein the step (B) comprises a step of:
 (B)(1) for each of a plurality of segments S of a spoken audio stream, performing steps of:
 (a) recognizing segment S with at least two of the plurality of probabilistic language models to identify at least two candidate contents for segment S; 
 (b) selecting one of the at least two candidate contents as final content corresponding to segment S; and 
 (c) inserting the final content for segment S into the document sub-structure associated with the probabilistic language model which produced the candidate content selected in step (B)(1)(b). 
   
     
     
         3 . The method of  claim 2 , wherein step (B)(1)(b) comprises steps of:
 (i) applying a metric to the at least two candidate contents to produce fitness scores for the at least two candidate contents, the fitness scores representing probabilities that the candidate contents represent the spoken audio stream; and   (ii) selecting the candidate content having the highest fitness score.   
     
     
         4 . The method of  claim 3 , wherein the step (B)(1)(b) further comprises a step of:
 (iii) selecting the probabilistic language model that was used to generate the candidate content selected in step (B)(1)(b)(ii).   
     
     
         5 . The method of  claim 1 , wherein the plurality of probabilistic language models includes at least one n-gram language model. 
     
     
         6 . The method of  claim 1 , wherein the plurality of probabilistic language models includes at least one finite state language model. 
     
     
         7 . The method of  claim 6 , wherein the plurality of probabilistic language models includes at least one n-gram language model. 
     
     
         8 . The method of  claim 1 , wherein the plurality of sub-structures of the document comprises a plurality of sections of the document. 
     
     
         9 . The method of  claim 1 , wherein the plurality of sections comprises a plurality of paragraphs. 
     
     
         10 . The method of  claim 1 , wherein the plurality of sub-structures includes a sub-structure representing a semantic concept. 
     
     
         11 . The method of  claim 10 , wherein the semantic concept comprises a date. 
     
     
         12 . The method of  claim 10 , wherein the semantic concept comprises a medication. 
     
     
         13 . The method of  claim 10 , wherein the semantic concept is represented in the document in a computer-readable form. 
     
     
         14 . The method of  claim 1 , further comprising a step of:
 (C) rendering the document to produce a rendition indicating the structure of the document.   
     
     
         15 . The method of  claim 1 , wherein the plurality of probabilistic language models are organized in a hierarchy, and wherein the step (B) comprises steps of:
 (B)(1) identifying a path through the hierarchy;   (B)(2) generating a document having a structure corresponding to the path identified in step (B)(1).   
     
     
         16 . The method of  claim 15 , wherein the step (B)(1) comprises a step of identifying a path through the hierarchy which, when applied by the speech recognition decoder to recognize the spoken audio stream, produces an optimal recognition result with respect to the hierarchy of the plurality of probabilistic language models. 
     
     
         17 . The method of  claim 15 , wherein step (B)(1) comprises steps of:
 (B)(1)(a) identifying a plurality of paths through the hierarchy;   (B)(1)(b) for each of the plurality of paths P, producing a candidate structured document for the spoken audio stream by using the speech recognition decoder to recognize the spoken audio stream using the language models on path P;   (B)(1)(c) applying a metric to the plurality of candidate structured documents produced in step (B)(1)(b) to produce a plurality of fitness scores for the plurality of candidate structured documents; and   (B)(1)(d) selecting the path which produces the candidate structured document having the highest fitness score.   
     
     
         18 . The method of  claim 1 , wherein the speech recognition decoder includes a plurality of speech recognition decoders, and wherein the step (B) includes steps of:
 (B)(1) identifying a segment of the spoken audio stream;   (B)(2) identifying one of the plurality of probabilistic language models;   (B)(3) identifying one of the plurality of speech recognition decoders having an association with the identified one of the plurality of probabilistic language models; and   (B)(4) using the identified speech recognition decoder to apply the identified probabilistic language model to the identified segment to produce content.   
     
     
         19 . The method of  claim 18 , wherein the identified one of the plurality of probabilistic language models comprises an n-gram language model, and wherein the identified speech recognition decoder comprises an n-gram speech recognition decoder. 
     
     
         20 . The method of  claim 18 , wherein the identified one of the plurality of probabilistic language models comprises a context-free grammar, and wherein the identified speech recognition decoder comprises a context-free grammar speech recognition decoder. 
     
     
         21 . The method of  claim 1 , wherein step (B) comprises steps of:
 (B)(1) identifying a mapping between the plurality of probabilistic language models and a plurality of segments in the audio stream;   (B)(2) for each of the plurality of segments, performing steps of:
 (B)(2)(a) identifying a corresponding one of the plurality of probabilistic language models using the mapping; 
 (B)(2)(b) identifying one of the plurality of sub-structures associated with the identified probabilistic language model; 
 (B)(3)(b) using the speech recognition decoder to recognize the segment using the identified one of the probabilistic language models thereby to produce content in the identified sub-structure. 
   
     
     
         22 . The method of  claim 21 , wherein the steps (B)(1) and (B)(2) are performed at least in part concurrently. 
     
     
         23 . The method of  claim 1 , wherein the step (B) comprises steps of:
 (B)(1) identifying a portion of the spoken audio stream representing semantic information; and   (B)(2) storing a representation of the semantic information in the document in a machine-readable form.   
     
     
         24 . An apparatus comprising:
 first identification means for identifying a probabilistic language model including a plurality of probabilistic language models associated with a plurality of sub-structures of a document;   document production means for using a speech recognition decoder to apply the probabilistic language model to a spoken audio stream to produce a document including content organized into the plurality of sub-structures, wherein the content in each of the plurality of sub-structures is produced by recognizing speech using the probabilistic language model associated with the sub-structure.   
     
     
         25 . The apparatus of  claim 24 , wherein the document production means comprises:
 iteration means comprising, for each of a plurality of segments S of a spoken audio stream:
 recognition means for recognizing segment S with at least two of the plurality of probabilistic language models to identify at least two candidate contents for segment S; 
 first selection means for selecting one of the at least two candidate contents as final content corresponding to segment S; and 
 insertion means for inserting the final content for segment S into the document sub-structure associated with the probabilistic language model which produced the candidate content selected by the selection means. 
   
     
     
         26 . The apparatus of  claim 25 , wherein the first selection means comprises:
 means for applying a metric to the at least two candidate contents to produce fitness scores for the at least two candidate contents, the fitness scores representing probabilities that the candidate contents represent the spoken audio stream; and   second selection means for selecting the candidate content having the highest fitness score.   
     
     
         27 . The apparatus of  claim 26 , wherein the first selection means further comprises:
 third selection means for selecting the probabilistic language model that was used to generate the candidate content selected by the second selection means.   
     
     
         28 . The apparatus of  claim 24 , wherein the plurality of probabilistic language models includes at least one n-gram language model. 
     
     
         29 . The apparatus of  claim 24 , wherein the plurality of probabilistic language models includes at least one finite state language model. 
     
     
         30 . The apparatus of  claim 29 , wherein the plurality of probabilistic language models includes at least one n-gram language model. 
     
     
         31 . The apparatus of  claim 24 , wherein the plurality of sub-structures of the document comprises a plurality of sections of the document. 
     
     
         32 . The apparatus of  claim 24 , wherein the plurality of sections comprises a plurality of paragraphs. 
     
     
         33 . The apparatus of  claim 24 , wherein the plurality of sub-structures includes a sub-structure representing a semantic concept. 
     
     
         34 . The apparatus of  claim 33 , wherein the semantic concept comprises a date. 
     
     
         35 . The apparatus of  claim 33 , wherein the semantic concept comprises a medication. 
     
     
         36 . The apparatus of  claim 33 , wherein the semantic concept is represented in the document in a computer-readable form. 
     
     
         37 . The apparatus of  claim 24 , further comprising:
 means for rendering the document to produce a rendition indicating the structure of the document.   
     
     
         38 . The apparatus of  claim 24 , wherein the plurality of probabilistic language models are organized in a hierarchy, and wherein the document production means comprises:
 second identification means for identifying a path through the hierarchy; and   means for generating a document having a structure corresponding to the path identified by the second identification means.   
     
     
         39 . The apparatus of  claim 38 , wherein the second identification means comprises means for identifying a path through the hierarchy which, when applied by the speech recognition decoder to recognize the spoken audio stream, produces an optimal recognition result with respect to the hierarchy of the plurality of probabilistic language models. 
     
     
         40 . The apparatus of  claim 38 , wherein the second identification means comprises:
 means for identifying a plurality of paths through the hierarchy;   candidate production means for producing, for each of the plurality of paths P, a candidate structured document for the spoken audio stream by using the speech recognition decoder to recognize the spoken audio stream using the language models on path P;   means for applying a metric to the plurality of candidate structured documents produced by the candidate production means to produce a plurality of fitness scores for the plurality of candidate structured documents; and   means for selecting the path which produces the candidate structured document having the highest fitness score.   
     
     
         41 . The apparatus of  claim 24 , wherein the speech recognition decoder includes a plurality of speech recognition decoders, and wherein the document production means includes:
 means for identifying a segment of the spoken audio stream;   means for identifying one of the plurality of probabilistic language models;   means for identifying one of the plurality of speech recognition decoders having an association with the identified one of the plurality of probabilistic language models; and   means for using the identified speech recognition decoder to apply the identified probabilistic language model to the identified segment to produce content.   
     
     
         42 . The apparatus of  claim 41 , wherein the identified one of the plurality of probabilistic language models comprises an n-gram language model, and wherein the identified speech recognition decoder comprises an n-gram speech recognition decoder. 
     
     
         43 . The apparatus of  claim 41 , wherein the identified one of the plurality of probabilistic language models comprises a context-free grammar, and wherein the identified speech recognition decoder comprises a context-free grammar speech recognition decoder. 
     
     
         44 . The apparatus of  claim 24 , wherein the document production means comprises:
 second identification means for identifying a mapping between the plurality of probabilistic language models and a plurality of segments in the audio stream;   iteration means comprising, for each of the plurality of segments:
 means for identifying a corresponding one of the plurality of probabilistic language models using the mapping; 
 means for identifying one of the plurality of sub-structures associated with the identified probabilistic language model; and 
 means for using the speech recognition decoder to recognize the segment using the identified one of the probabilistic language models thereby to produce content in the identified sub-structure. 
   
     
     
         45 . The apparatus of  claim 44 , wherein the second identification means and the iteration means are configured to operate at least in part concurrently. 
     
     
         46 . The apparatus of  claim 24 , wherein the document production means comprises:
 means for identifying a portion of the spoken audio stream representing semantic information; and   means for storing a representation of the semantic information in the document in a machine-readable form.   
     
     
         47 . A data structure comprising the language model identified in step (A) of  claim 1 . 
     
     
         48 . The data structure of  claim 47 , wherein the plurality of probabilistic language models includes at least one n-gram language model. 
     
     
         49 . The data structure of  claim 47 , wherein the plurality of probabilistic language models includes at least one context-free grammar. 
     
     
         50 . The data structure of  claim 49 , wherein the plurality of probabilistic language models includes at least one n-gram language model. 
     
     
         51 . A data structure comprising:
 a plurality of language models logically organized in a hierarchy, the plurality of language models including a first language model and a second language model;   wherein the first language model is a parent of the second language model in the hierarchy;   wherein the first language model is suitable for recognizing speech representing a first concept associated with a substructure of a document; and   wherein the second language model is suitable for recognizing speech representing a second concept associated with a subset of the substructure of the document.   
     
     
         52 . The data structure of  claim 51 , wherein the first language model comprises an n-gram language model. 
     
     
         53 . The data structure of  claim 51 , wherein the first language model comprises a context-free grammar. 
     
     
         54 . The data structure of  claim 53 , wherein the second language model comprises an n-gram language model. 
     
     
         55 . The data structure of  claim 51 , wherein the substructure of the document comprises a section of the document.