Semantic representation and realization for conversational systems

Patent metadata

Field	Value
Publication number	US-10845937-B2
Application number	US-201815868987-A
Country	US
Kind code	B2
Filing date	Jan 11, 2018
Priority date	Jan 11, 2018
Publication date	Nov 24, 2020
Grant date	Nov 24, 2020

Abstract

A method, apparatus and computer program product for presenting a user interface for a conversational system is described. A unified semantic representation of conversational content between a user and conversational system is created as a contextual graph of concepts and relations. A set of subgraph components of the semantic contextual graph dynamically identified based on a current dialog activity. The identified set of subgraph components in a user interface as a set of graphical elements representing respective concepts and relations.

First claim

What is claimed:

1. A method for presenting a user interface for a conversational system comprising:
creating a unified semantic representation of conversational content between a user and conversational system as a semantic contextual graph of concepts and relations, wherein the unified semantic representation contains conversational content from both the conversational system and the user from a current conversation;
dynamically identifying a set of subgraph components of the semantic contextual graph to be presented to the user based on use of the set of subgraph components to create a conversational system response in a current dialog activity in the current conversation;
rendering the identified set of subgraph components in the user interface as a set of graphical elements representing respective concepts and relations, wherein the identified set of graphical elements accepts user input:
receiving a first user input directed to the respective graphical element;
providing the first user input to a graph integrator component of the conversational system which produces a meaning representation graph corresponding to the first user input; and
changing the unified semantic representation based on the meaning representation graph;
wherein each graphical element of the identified set of graphical elements is displayed in a manner indicating an origin of a corresponding subgraph component of the unified semantic representation, wherein the indicating an origin includes indicating that respective subgraph components of the unified semantic representation originate from the conversational content from both the conversational system and the user from the current conversation, from an original ontology which preexisted the current conversation and from a query to a backend database coupled to the unified semantic representation.

2. The method as recited in claim 1 , wherein the unified semantic representation contains subgraphs representing system information sources which existed prior to the current conversation and at least one of the identified set of subgraph components rendered in the user interface is based on the system information sources which existed prior to the current conversation.

3. The method as recited in claim 2 , further comprising:
dynamically identifying a plurality of subgraph components of the semantic contextual graph relevant to the current dialog activity;
identifying a set of constraints that prevents the entire identified set of subgraph components from being presented to the user as a corresponding plurality of graphical elements, only allowing a subset of the identified set to be presented; and
optimizing the identified set of subgraph components from the plurality of subgraph components based on a semantic completeness criterion to help end users to comprehend the set of graphical elements in the user interface.

4. The method as recited in claim 3 , wherein the constraints are a set of temporal and spatial constraints for rendering the identified set of subgraph components in the user interface, and the method further comprises:
identifying the subgraph components that correspond to a likelihood of the user providing feedback if the respective graphical elements are presented representing respective subgraph components;
as a spatial constraint, estimating a display space required for rendering respective graphical elements as compared to a total available space for the set of graphical elements components in the user interface; and
as a temporal constraint, estimating a time required for rendering respective graphical elements in the user interface as compared to a recent pace of the current conversation.

5. The method as recited in claim 1 , wherein the set of graphical elements representing respective concepts and relations is user annotatable so that user feedback to the conversational system is supplied from user interaction with the set of graphical elements.

6. The method as recited in claim 1 , further comprising responsive to subgraph components which correspond to the current dialog activity being unavailable, scoring a set of candidate subgraph components according to how semantically related is a respective candidate subgraph component to a current user intent and how likely the user will provide feedback if the respective subgraph component is displayed;
wherein the candidate subgraph components are obtained by a predetermined number of hops away from a concept node which represents the current user intent in the unified semantic representation.

7. The method as recited in claim 1 , wherein the identified set of subgraph components are identified according to an expertise level of the user.

8. The method as recited in claim 1 , wherein the respective subgraph components of the unified semantic representation that originate from the conversational content from both the conversational system and the user from the current conversation, from an original ontology which preexisted the current conversation and from a query to a backend database coupled to the unified semantic representation are respectively displayed in different colors.

9. Apparatus, comprising:
a processor;
computer memory holding computer program instructions executed by the processor for searching for presenting a user interface for a conversational system, the computer program instructions comprising:
program code, operative to create a unified semantic representation of conversational content between a user and conversational system as a semantic contextual graph of concepts and relations, wherein the unified semantic representation contains conversational content from both the conversational system and the user from a current conversation;
program code operative to dynamically identify a set of subgraph components of the semantic contextual graph to be presented to the user based on use of the set of subgraph components to create a system response in a current dialog activity in the current conversation; and
program code operative to render the identified set of subgraph components in the user interface as a set of graphical elements representing respective concepts and relations, wherein the set of graphical elements accept user input:
program code, operative to receive a first user input directed to the respective graphical element;
program code, operative to provide the first user input to a graph integrator component of the conversational system which produces a meaning representation graph corresponding to the first user input; and
program code, operative to change the unified semantic representation based on the meaning representation graph;
wherein each graphical element of the identified set of graphical elements is displayed in a manner indicating an origin of a corresponding subgraph component of the unified semantic representation, wherein the indicating an origin includes indicating that respective subgraph components of the unified semantic representation originate from the conversational content from both the conversational system and the user from the current conversation, from an original ontology which preexisted the current conversation and from a query to a backend database coupled to the unified semantic representation.

10. The apparatus as recited in claim 9 , wherein the unified semantic representation contains subgraphs representing system information sources which existed prior to the current conversation and at least one of the identified set of subgraph components rendered in the user interface is based on the system information sources which existed prior to the current conversation.

11. The apparatus as recited in claim 9 , further comprising: prog

Assignees

• IBM

Inventors

• Anand Rangachari
• Arora Ashima
• Bakis Raimo
• Feng Song
• Ganhotra Jatin
• Gunasekara Chulaka
• Nahamoo David
• Polymenakos Lazaros
• Shashidhara Sunil D
• Zhu Li

Classifications

• G06F9/451 Primary

  Execution arrangements for user interfaces · CPC title

• G06Q50/20

  Education · CPC title

• G06Q50/10

  Services · CPC title

• G06Q10/10

  Office automation; Time management · CPC title

• G06F3/04842

  Selection of displayed objects or displayed text elements (G06F3/0482 takes precedence) · CPC title