Semantic representation and realization for conversational systems

Abstract

A method, apparatus and computer program product for presenting a user interface for a conversational system is described. For each of a set of user utterances produced in a dialog with the conversational system, a semantic meaning representation is determined. The semantic meaning representations are converted to respective sentential concept graphs. A first sentential concept graph is consolidated into a unified contextual graph. The unified contextual graph is updated based on new sentential concept graphs while the dialog with the conversational system progresses.

What is claimed: 
     
         1 . A method for presenting a user interface for a conversational system comprising:
 for each of a set of user utterances produced in a dialog with the conversational system, determining a semantic meaning representation;   converting the semantic meaning representations to respective sentential concept graphs;   consolidating a first sentential concept graph into a unified contextual graph; and   updating the unified contextual graph based on new sentential concept graphs while the dialog with the conversational system progresses.   
     
     
         2 . The method as recited in  claim 1  further comprising updating the unified contextual graph based on semantic matching with domain knowledge stored in a database of the conversational system. 
     
     
         3 . The method as recited in  claim 1 , further comprising dynamically identifying content from the united contextual graph to present in a user interface. 
     
     
         4 . The method as recited in  claim 3 , further comprising identifying concepts and relations that are semantically related to a latest dialog activity in the dialog with the conversational system. 
     
     
         5 . The method as recited in  claim 3 , further comprising:
 identifying concepts and relations that are semantically related in the first sentential concept graph to concepts and relations in the unified contextual graph; and   constructing a query to a database of the conversational system according to the identified concepts and relations.   
     
     
         6 . The method as recited in  claim 3 , further comprising;
 identifying a set of changes to concepts, concept values and concept statuses based on a latest user input;   identifying relevant components to the concepts, the concept values and the concept statuses in the unified contextual graph; and   identifying changes to relevant components in the unified contextual graph based on results from queries to databases of the conversational system.   
     
     
         7 . The method as recited in  claim 6 , further comprising:
 quantifying a set of factors associated with respective subgraphs, the set of factors including at least one of a concept-level feature, a relation-level feature or a discourse-level feature; and   ranking a set of subgraphs in the unified contextual graph based on the quantified factors.   
     
     
         8 . The method as recited in  claim 3 , further comprising generating a dialog policy according to a new user intent and a historical utterance in the dialog rather than a latest utterance in the dialog. 
     
     
         9 . Apparatus, comprising:
 a processor;   computer memory holding computer program instructions executed by the processor for presenting a user interface for a conversational system, the computer program instructions comprising:
 program code, operative to determine a semantic meaning representation for each of a set of user utterances produced in a dialog with the conversational system; 
 program code, operative to convert the semantic meaning representations to respective sentential concept graphs; 
 program code, operative to consolidate a first sentential concept graph into a unified contextual graph; and 
 program code, operative to update the unified contextual graph based on new sentential concept graphs while the dialog with the conversational system progresses. 
   
     
     
         10 . The apparatus as recited in  claim 9 , further comprising:
 program code, operative to dynamically identify content from the united contextual graph to present in a user interface.   
     
     
         11 . The apparatus as recited in  claim 10 , further comprising:
 program code, operative to identify concepts and relations that are semantically related to a latest dialog activity in the dialog with the conversational system.   
     
     
         12 . The apparatus as recited in  claim 10 , further comprising:
 program code, operative to identify concepts and relations that are semantically related in the first sentential concept graph to concepts and relations in the unified contextual graph; and   program code, operative to construct a query to a database of the conversational system according to the identified concepts and relations.   
     
     
         13 . The apparatus as recited in  claim 10 , further comprising:
 program code, operative to quantify a set of factors associated with respective subgraphs, the set of factors including at least one of a concept-level feature, a relation-level feature or a discourse-level feature; and   program code, operative to rank a set of subgraphs in the unified contextual graph based on the quantified factors.   
     
     
         14 . The apparatus as recited in  claim 10 , further comprising:
 program code, operative to generate a dialog policy according to a new user intent and a historical utterance in the dialog rather than a latest utterance in the dialog.   
     
     
         15 . A computer program product in a non-transitory computer readable medium for use in a data processing system, the computer program product holding computer program instructions executed by the data processing system for presenting a user interface for a conversational system, the computer program instructions comprising:
 program code, operative to determine a semantic meaning representation for each of a set of user utterances produced in a dialog with the conversational system;   program code, operative to convert the semantic meaning representations to respective sentential concept graphs;   program code, operative to consolidate a first sentential concept graph into a unified contextual graph; and   program code, operative to update the unified contextual graph based on new sentential concept graphs while the dialog with the conversational system progresses   
     
     
         16 . The computer program product as recited in  claim 15 , further comprising:
 program code, operative to dynamically identify content from the united contextual graph to present in a user interface.   
     
     
         17 . The computer program product as recited in  claim 16 , further comprising:
 program code, operative to identify concepts and relations that are semantically related to a latest dialog activity in the dialog with the conversational system.   
     
     
         18 . The computer program product as recited in  claim 16 , further comprising:
 program code, operative to identify concepts and relations that are semantically related in the first sentential concept graph to concepts and relations in the unified contextual graph; and   program code, operative to construct a query to a database of the conversational system according to the identified concepts and relations.   
     
     
         19 . The computer program product as recited in  claim 15 , wherein 
     
     
         20 . The computer program product as recited in  claim 16 , further comprising:
 program code, operative to generate a dialog policy according to a new user intent and a historical utterance in the dialog rather than a latest utterance in the dialog.